让新闻交易轻松上手（第二部分）：风险管理·综合运用
- 把多国夏令时塞进一个基类
- 用类封装夏令时区间判定
- 夏令时切换下的事件时间补偿
- 澳洲夏令时切换在回测里怎么落地
- 澳洲夏令时切换的硬编码区间
- 英国夏令时切换的回测校准类
- 把英国夏令时区间写死进列表
- 把美国夏令时写进回测时钟
- 美式夏令时切换的硬编码清单
- 把交易品种属性收进一个类里
- 封装品种属性的类骨架
- 封装品种属性读取的 C++ 类接口
- 把合约规格和手数限制一次性读出来
- 抓点差与止损冻结位的类方法
- 封装品种属性读取的容错逻辑
- 挂单与持仓量的类内抓取逻辑
- 读取品种保证金与自定义状态的取值细节
- 把品种属性封装成可复用取数函数
- 封装品种交易限制与报价的取数接口
- 从品种属性类里抠出实时盘口数据
- 从交易品种扒出币种与背景色
- 时间操作类的接口设计
- 时间类的秒数换算与区间判定
- 时间管理类的年月日与偏移拆解
- EA卸载后自动还原图表外观
- 用类把图表属性一次性存回原样
- 用类封装一键套用图表外观
- 用类封装把图表刷成黑白极简风
- 把图表外观塞进一个结构体
- 用类构造函数批量接管图表属性
- 用类封装把图表属性一次性写进MT5
- 用代码接管图表外观与析构还原
- K线属性类的继承与取价接口
- 用K线高度比对锁定异动发生柱
- 图表对象的创建与回收机制
- 用矩形标签在图表上画可控色块
- 用 OBJ_LABEL 在图表上钉死文字坐标
- 在图表上动态画事件标记与方框
- 用矩形标签铺自定义面板底图
- 用代码在图表上钉死标签和事件锚点
- 把重复新闻拆成两张表省一半体积
- 用 SQLITE_MASTER 摸清库里到底有哪些表
- 看一眼本地库表结构
- 给财经日历库里的对象发身份证
- 财经日历类的结构骨架与取数函数
- 用视图把财经日历按时区拆开
- 日历库的三张表与单例触发器
- 用触发器锁死单条记录与抓取财经日历
- 把财经日历事件塞进自定义结构体
- 用 sqlite_master 给财经库做结构对账
- 用 SQL 比对给财经日历库做增量更新判定
- 经纪商财经日历库的本地建表逻辑
- 建表失败时的回滚与成功后的视图落地
- 建表前的删表兜底与失败退出
- 把财经日历落库时顺手调好英美夏令时
- 夏令时校正与更新记录表的落地写法
- 把经纪商夏令时写进本地库的事务兜底
- 从财经库里抠出最近更新时间
- 抓去年美国NFP日期并定位经纪商EURUSD
- 用 EURUSD 异动反推券商时区切换
- 用夏令时枚举给经纪商时区定性
- 用枚举和结构体接管经纪商 DST 表
- 按内容字段捞回单条财经日历
- SQLite里的PRAGMA到底管什么
- 用 PRAGMA 给日历库做干净的重置
- 经济日历原始字段怎么读
- 经济日历原始条目的字段排布
- 经济日历里的低权重条目与夏令时标记
- 历史成交记录的同时间戳堆叠现象
- 时间列对齐错位会骗过你的回测
- 用视图把财经日历按交易时段拆开
- 把财经日历拼成可查的视图
- 把财经日历条目落进本地库表
- 用触发器锁死单条配置表
- 用触发器清空旧记录再写库
- 清理冗余表并判定财经日历库是否需更新
- 建表前先清旧结构的坑
- 财经日历落库的函数骨架
- 夏令时校正后落库的事件时间
- 服务器时间才是跨品种记录的基准
- 自动夏令时触发的建表与回滚兜底
- 手数风险档位怎么切
- 用余额百分比反推下单手数
- 未平仓时为何看可用保证金占比
- 按余额比例折算每笔风险金
- 按可用保证金切风险额的算法
- 按余额比例算手数会被步长卡住
- 按可用保证金倒推开仓手数
- 用保证金反推最大可扛风险手数
- 风控模块的账户基准与仓位结构
- 风控类里的手数推导与订单类型判定
- 仓位风险档位的字符串映射
- 风险档位如何映射成手数返回值
- 仓位计算的分支出口与风险函数实现
- 手数封顶与步进对齐的兜底逻辑
- 手数下限与上限的硬约束逻辑
- 手数封顶的多档熔断逻辑
- 把风险档位映射成可读字符串
- 风险上下限的字符串映射与初始化
- 手数决策的 switch 分支与风险映射
- 三种风险档位与手数下限校验
- 手数封顶与品种限仓的兜底逻辑
- 手数触顶时的降级与清零逻辑
- 通用图形类如何接管品种属性与风险面板
- 把品种手数约束画到图表左上角
- 把风控参数钉在图表左上角
- 把新闻EA挂上图表前的初始化动作
- 风控档位的初始化边界
- 断网时的等待动画怎么画
- 初始化时的数据库就绪检查
- 手算仓位与风险类型在初始化里的落地
- 别急着下结论
把多国夏令时塞进一个基类
做跨市场外汇或贵金属回测时，英美澳三地的夏令时切换日不一致，直接拿本地时间算 session 会偏掉 1 小时。把共性抽成 CDaylightSavings 单一继承 CObject，后续各国子类只重写虚函数填自己的时间表，主逻辑不用改。 类里挂了 CTimeManagement 实例和两组 CArrayObj 指针（savings / getSavings），分别存已初始化表和查询缓冲。保护段声明了 SetDaylightSavings_UK / US / AU 三个虚函数，就是给子类留的坑位。 对外暴露的 isDaylightSavings(datetime) 判某时刻是否落在夏令时内，DaylightSavings(int Year,...) 按年取起止日，adjustDaylightSavings(datetime) 按当时区规则把事件时间做时区折算。外汇与贵金属杠杆高、跳空频繁，时区算错可能把新闻交易触发器开在错的时间点，实盘前务必在 MT5 用历史数据跑一遍这三个函数。
class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| NewsTrading | class=class="str">"cmt">//| Copyright class="num">2024, MetaQuotes Ltd. | class=class="str">"cmt">//| [MQL5官方文档] | class=class="str">"cmt">//+------------------------------------------------------------------+ class="macro">#include <Object.mqh> class="macro">#include <Arrays\ArrayObj.mqh> class="macro">#include "../TimeManagement.mqh" class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|DaylightSavings class | class=class="str">"cmt">//+------------------------------------------------------------------+ class CDaylightSavings: class="kw">public CObject { class="kw">protected: CTimeManagement Time; CDaylightSavings(class="type">class="kw">datetime startdate,class="type">class="kw">datetime enddate); CObject *List() { class="kw">return savings;}\n class="type">class="kw">datetime StartDate; class="type">class="kw">datetime EndDate; CArrayObj *savings; CArrayObj *getSavings; CDaylightSavings *dayLight; class="kw">virtual class="type">void SetDaylightSavings_UK(); class="kw">virtual class="type">void SetDaylightSavings_US(); class="kw">virtual class="type">void SetDaylightSavings_AU(); class="kw">public: CDaylightSavings(class="type">void); ~CDaylightSavings(class="type">void); class="type">bool isDaylightSavings(class="type">class="kw">datetime Date); class="type">bool DaylightSavings(class="type">int Year,class="type">class="kw">datetime &startDate,class="type">class="kw">datetime &endDate); class="type">class="kw">string adjustDaylightSavings(class="type">class="kw">datetime EventDate); };
「用类封装夏令时区间判定」
在 MT5 写跨时区策略时，夏令时切换会让 K 线时间戳偏移 1 小时，手动硬改偏移量容易在每年 3 月、11 月踩坑。把夏令时逻辑收进一个 CDaylightSavings 类，比在 EA 主循环里散落判断要干净得多。 类提供了两个构造器：无参版本只做空初始化，带 startdate 与 enddate 的版本直接把起止时间赋给类内全局变量 StartDate、EndDate。这样你在回测不同年份时，可以显式传入该年夏令时窗口。 isDaylightSavings() 方法先 getSavings = List() 拉出已存的夏令时段列表，再 for 循环遍历，用 Time.DateIsInRange(dayLight.StartDate, dayLight.EndDate, Date) 判断传入日期是否落在某一段内，命中即 return true，全遍历无命中则返回 false。 DaylightSavings(Year, startDate, endDate) 则按年检索：用 Time.ReturnYear() 比对列表中每段的起、止年份，分别置 startDateDetected / endDateDetected 标志，两者皆真才认为该年数据齐备。外汇与贵金属受夏令时影响跳空概率偏高，用这类封装可降人工配错风险。 别把 List() 当黑盒 代码里 getSavings = List() 每次调用都重新初始化，若列表从外部文件或接口载入，应在构造时缓存而非每次判段重建，否则 112 节后续高频调用会拖慢 tick 处理。
class=class="str">"cmt">//|Constructor | class=class="str">"cmt">//+------------------------------------------------------------------+ CDaylightSavings::CDaylightSavings(class="type">void) { } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Initialize variables | class=class="str">"cmt">//+------------------------------------------------------------------+ CDaylightSavings::CDaylightSavings(class="type">class="kw">datetime startdate,class="type">class="kw">datetime enddate) { StartDate = startdate;class=class="str">"cmt">//Assign class&class="macro">#x27;s global variable StartDate value from parameter variable startdate EndDate = enddate;class=class="str">"cmt">//Assign class&class="macro">#x27;s global variable EndDate value from parameter variable enddate } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|checks if a given date falls within Daylight Savings Time | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">bool CDaylightSavings::isDaylightSavings(class="type">class="kw">datetime Date) { class=class="str">"cmt">// Initialize a list to store daylight savings periods. getSavings = List(); class=class="str">"cmt">// Iterate through all the periods in the list. for(class="type">int i=class="num">0; i<getSavings.Total(); i++) { class=class="str">"cmt">// Access the current daylight savings period. dayLight = getSavings.At(i); class=class="str">"cmt">// Check if the given date is within the current daylight savings period. if(Time.DateIsInRange(dayLight.StartDate,dayLight.EndDate,Date)) { class=class="str">"cmt">// If yes, class="kw">return true indicating it is daylight savings time. class="kw">return true; } } class=class="str">"cmt">// If no period matches, class="kw">return false indicating it is not daylight savings time. class="kw">return false; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Check if DaylightSavings Dates are available for a certain Year | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">bool CDaylightSavings::DaylightSavings(class="type">int Year,class="type">class="kw">datetime &startDate,class="type">class="kw">datetime &endDate) { class=class="str">"cmt">// Initialize a list to store daylight savings periods. getSavings = List(); class="type">bool startDateDetected=false,endDateDetected=false; class=class="str">"cmt">// Iterate through all the periods in the list. for(class="type">int i=class="num">0; i<getSavings.Total(); i++) { dayLight = getSavings.At(i); if(Year==Time.ReturnYear(dayLight.StartDate))class=class="str">"cmt">//Check if a certain year&class="macro">#x27;s date is available within the DaylightSavings start dates in the List { startDate = dayLight.StartDate; startDateDetected = true; } if(Year==Time.ReturnYear(dayLight.EndDate))class=class="str">"cmt">//Check if a certain year&class="macro">#x27;s date is available within the DaylightSavings end dates in the List { endDate = dayLight.EndDate; endDateDetected = true; } if(startDateDetected&&endDateDetected)class=class="str">"cmt">//Check if both DaylightSavings start and end dates are found for a certain Year {
◍ 夏令时切换下的事件时间补偿
这段逻辑处理的是交易服务器时间与事件时间处于不同夏令时状态时的偏移补偿。核心判断来自 isDaylightSavings(TimeTradeServer()) 与 isDaylightSavings(EventDate) 的组合，两者不一致时就需要对 EventDate 做 ±1 小时修正。 若服务器已在夏令时内，而事件时间不在，则返回 TimeToString((datetime)(EventDate - Time.HoursS()))，相当于把事件时间往前推 1 小时（HourS() 默认返回 3600 秒）。反过来，服务器不在夏令时、事件在夏令时，则加 1 小时：TimeToString((datetime)(Time.HoursS() + EventDate))。 默认兜底写在查找失败处：startDate 与 endDate 都被设为 D'1970.01.01 00:00:00'，即 Unix 纪元零点，调用方需自行识别这个哨兵值，避免把 1970 年当成真实夏令时区间。 析构函数里连续 delete 了 savings、dayLight、getSavings 三个指针，说明该类内部用 CArrayObj 和自身实例做了组合持有，漏掉任一 delete 都会在 EA 反复加载时泄漏内存。
class="kw">return true; } } startDate = D&class="macro">#x27;class="num">1970.01.class="num">01 class="num">00:class="num">00:class="num">00&class="macro">#x27;;class=class="str">"cmt">//Set a class="kw">default start date if no DaylightSaving date is found endDate = D&class="macro">#x27;class="num">1970.01.class="num">01 class="num">00:class="num">00:class="num">00&class="macro">#x27;;class=class="str">"cmt">//Set a class="kw">default end date if no DaylightSaving date is found class="kw">return false; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Will adjust the date&class="macro">#x27;s timezone depending on DaylightSavings | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">string CDaylightSavings::adjustDaylightSavings(class="type">class="kw">datetime EventDate) { if(isDaylightSavings(TimeTradeServer()))class=class="str">"cmt">//Check if the current tradeserver time is already within the DaylightSavings Period { if(isDaylightSavings(EventDate))class=class="str">"cmt">//Checks if the event time is during daylight savings { class="kw">return TimeToString(EventDate);class=class="str">"cmt">//normal event time } else { class="kw">return TimeToString((class="type">class="kw">datetime)(EventDate-Time.HoursS()));class=class="str">"cmt">//event time minus an hour for DST } } else { if(isDaylightSavings(EventDate))class=class="str">"cmt">//Checks if the event time is during daylight savings { class="kw">return TimeToString((class="type">class="kw">datetime)(Time.HoursS()+EventDate));class=class="str">"cmt">//event time plus an hour for DST } else { class="kw">return TimeToString(EventDate);class=class="str">"cmt">//normal event time } } } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Destructor | class=class="str">"cmt">//+------------------------------------------------------------------+ CDaylightSavings::~CDaylightSavings(class="type">void) { class="kw">delete savings;class=class="str">"cmt">//Delete CArrayObj Pointer class="kw">delete dayLight;class=class="str">"cmt">//Delete CDaylightSavings Pointer class="kw">delete getSavings;class=class="str">"cmt">//Delete CArrayObj Pointer } class=class="str">"cmt">//+------------------------------------------------------------------+
澳洲夏令时切换在回测里怎么落地
做澳元系货币对或 XAU/AUD 的跨年回测时，若直接用 MT5 策略测试器默认时区，夏令时缺口会让 4 月与 10 月的 K 线时间戳偏移，导致新闻触发逻辑误判。CDaylightSavings_AU 类就是专门把澳洲历年夏令时起止写死进数据库，让测试器重排时间。 该类公开继承 CDaylightSavings，后者又挂在 CObject 与 CArrayObj 之上，本质是一个用对象数组承载时间段的结构。构造函数只留空壳，真正的日程填充发生在 SetDaylightSavings_AU 这个虚函数重载里。 代码里 new 了一个 CArrayObj 叫 savings，然后逐条 Add 澳洲夏令时区间：例如 2006.10.29 03:00 进入、2007.03.25 02:00 退出，2007.10.28 03:00 进入、2008.04.06 02:00 退出，一直排到 2016.10.02 至 2017.04.02，共 11 个年度窗口。你打开 MT5 按这套日期核对自己日历，若做 2017 年后的澳系回测，需自行补写新年份，否则可能漏掉时差重排。 外汇与贵金属杠杆高、跳空频繁，夏令时错配会放大滑点风险，建议每次换季前手动校验该数组上界。
class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| NewsTrading | class=class="str">"cmt">//| Copyright class="num">2024, MetaQuotes Ltd. | class=class="str">"cmt">//| [MQL5官方文档] | class=class="str">"cmt">//+------------------------------------------------------------------+ class="macro">#include "DaylightSavings.mqh" class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|DaylightSavings_AU class | class=class="str">"cmt">//+------------------------------------------------------------------+ class CDaylightSavings_AU: class="kw">public CDaylightSavings { class="kw">public: CDaylightSavings_AU(class="type">void); }; class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Set Daylight Savings Schedule for Australia | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">void CDaylightSavings::SetDaylightSavings_AU() { savings = new CArrayObj(); class=class="str">"cmt">//Daylight savings dates to readjust dates in the database for accurate testing in the strategy tester savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2006.10.class="num">29 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2007.03.class="num">25 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2007.10.class="num">28 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2008.04.class="num">06 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2008.10.class="num">05 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2009.04.class="num">05 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2009.10.class="num">04 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2010.04.class="num">04 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2010.10.class="num">03 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2011.04.class="num">03 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2011.10.class="num">02 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2012.04.class="num">01 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2012.10.class="num">07 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2013.04.class="num">07 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2013.10.class="num">06 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2014.04.class="num">06 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2014.10.class="num">05 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2015.04.class="num">05 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2015.10.class="num">04 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2016.04.class="num">03 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2016.10.class="num">02 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2017.04.class="num">02 class="num">02:class="num">00:class="num">00&class="macro">#x27;));
「澳洲夏令时切换的硬编码区间」
在 MT5 里做跨时区行情对齐时，最省事但也最容易过期的办法，就是把每年澳洲夏令时的起止时间直接写死进一个 CDaylightSavings 列表。下面这段代码从 2017 年一路排到 2029 年，每年 10 月第一个周日 03:00 时钟拨快到夏令时，次年 4 月第一个周日 02:00 拨回标准时。 具体看几条典型记录：2017.10.01 03:00 进夏令、2018.04.01 02:00 退出；2024.10.06 03:00 进、2025.04.06 02:00 退。能明显看出起始日逐年在前移（10 月 1 日→10 月 7 日之间浮动），结束日则在 4 月 1 日→7 日之间摆动，这不是固定日期，而是按澳洲法定规则逐年推算后落表的。 构造函数 CDaylightSavings_AU 本身只做一件事：在实例化时调用 SetDaylightSavings_AU() 把上面那串区间灌进 savings 容器。你如果要在 EA 里判断当前是否处于夏令时段，直接拿这个类的实例去比对自己的本地时间即可。 外汇与贵金属受时区切换影响大，点差和流动性在切换周可能异常，实盘使用前建议先在策略测试器用 2019–2024 年数据回放一遍时间判定逻辑。
savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2017.10.class="num">01 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2018.04.class="num">01 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2018.10.class="num">07 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2019.04.class="num">07 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2019.10.class="num">06 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2020.04.class="num">05 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2020.10.class="num">04 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2021.04.class="num">04 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2021.10.class="num">03 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2022.04.class="num">03 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2022.10.class="num">02 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2023.04.class="num">02 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2023.10.class="num">01 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2024.04.class="num">07 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2024.10.class="num">06 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2025.04.class="num">06 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2025.10.class="num">05 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2026.04.class="num">05 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2026.10.class="num">04 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2027.04.class="num">04 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2027.10.class="num">03 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2028.04.class="num">02 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2028.10.class="num">01 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2029.04.class="num">01 class="num">02:class="num">00:class="num">00&class="macro">#x27;)); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Constructor | class=class="str">"cmt">//+------------------------------------------------------------------+ CDaylightSavings_AU::CDaylightSavings_AU(class="type">void) { SetDaylightSavings_AU(); } class=class="str">"cmt">//+------------------------------------------------------------------+
◍ 英国夏令时切换的回测校准类
在新闻交易类的 EA 架构里，CDaylightSavings_UK 负责把英国（欧洲）的夏令时起止时间单独抽成一个类，继承自 CDaylightSavings，上层还挂在 CObject、CArrayObj、CTimeManagement 这条链上。它的存在不是为了下单，而是给策略测试器喂准确的时区偏移，否则 2007–2017 这段历史里伦敦时制切换日被当成普通日，回测的欧美盘重叠时段会偏掉。 类里真正干活的是 SetDaylightSavings_UK 这个扩展方法：先 new 一个 CArrayObj 容器 savings，再把每年两次的切换点以 CDaylightSavings 对象压进去。比如 2007 年是 3 月 25 日 02:00 拨快、10 月 28 日 01:00 拨回；2017 年则是 3 月 26 日进夏令、10 月 29 日退夏令，规律是三月末周日进、十月末周日出。 你在 MT5 里接这套代码时，直接把 DaylightSavings.mqh 包含进来就能用。若你的样本要测 2018 之后，得自己按相同格式补 Add 行——原片段只列到 2017，缺了年份测试器就会静默跳过校准。外汇与贵金属受时区跳变影响报价间隙，属高风险品种，参数补全后建议先跑单年样本验证偏移量。
class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| NewsTrading | class=class="str">"cmt">//| Copyright class="num">2024, MetaQuotes Ltd. | class=class="str">"cmt">//| [MQL5官方文档] | class=class="str">"cmt">//+------------------------------------------------------------------+ class="macro">#include "DaylightSavings.mqh" class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|DaylightSavings_UK class | class=class="str">"cmt">//+------------------------------------------------------------------+ class CDaylightSavings_UK: class="kw">public CDaylightSavings { class="kw">public: CDaylightSavings_UK(class="type">void); }; class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Set Daylight Savings Schedule for Europe | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">void CDaylightSavings::SetDaylightSavings_UK() { savings = new CArrayObj(); class=class="str">"cmt">//Daylight savings dates to readjust dates in the database for accurate testing in the strategy tester savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2007.03.class="num">25 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2007.10.class="num">28 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2008.03.class="num">30 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2008.10.class="num">26 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2009.03.class="num">29 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2009.10.class="num">25 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2010.03.class="num">28 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2010.10.class="num">31 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2011.03.class="num">27 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2011.10.class="num">30 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2012.03.class="num">25 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2012.10.class="num">28 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2013.03.class="num">31 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2013.10.class="num">27 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2014.03.class="num">30 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2014.10.class="num">26 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2015.03.class="num">29 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2015.10.class="num">25 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2016.03.class="num">27 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2016.10.class="num">30 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2017.03.class="num">26 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2017.10.class="num">29 class="num">01:class="num">00:class="num">00&class="macro">#x27;));
把英国夏令时区间写死进列表
处理跨时区行情时，最怕夏令时切换日 K 线出现缺口或重复小时。上面这段把 2018–2029 年英国夏令时的起止点，用 CDaylightSavings 对象逐个塞进 savings 列表：起始都是当年 3 月最后一个周日的 02:00:00，结束都是 10 月最后一个周日的 01:00:00。 注意 2021 与 2027 的结束日是 10.31，而 2020 与 2026 是 10.25，差了将近一周，硬算公式容易错，直接枚举反而最稳。构造函数 CDaylightSavings_UK 里只调了 SetDaylightSavings_UK()，也就是把这张表在实例化时一次性加载好。 开 MT5 后把这段贴进 EA 或脚本，编译跑一遍，用 Print(savings.Size()) 应回吐 12，说明十二年区间已就绪。外汇与贵金属受时区跳变影响大，切换周报价可能异常，实盘前先在策略测试器验证。
savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2018.03.class="num">25 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2018.10.class="num">28 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2019.03.class="num">31 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2019.10.class="num">27 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2020.03.class="num">29 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2020.10.class="num">25 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2021.03.class="num">28 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2021.10.class="num">31 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2022.03.class="num">27 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2022.10.class="num">30 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2023.03.class="num">26 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2023.10.class="num">29 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2024.03.class="num">31 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2024.10.class="num">27 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2025.03.class="num">30 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2025.10.class="num">26 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2026.03.class="num">29 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2026.10.class="num">25 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2027.03.class="num">28 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2027.10.class="num">31 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2028.03.class="num">26 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2028.10.class="num">29 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2029.03.class="num">25 class="num">02:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2029.10.class="num">28 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Constructor | class=class="str">"cmt">//+------------------------------------------------------------------+ CDaylightSavings_UK::CDaylightSavings_UK(class="type">void) { SetDaylightSavings_UK(); } class=class="str">"cmt">//+------------------------------------------------------------------+
「把美国夏令时写进回测时钟」
做新闻类 EA 最容易被忽略的坑，是策略测试器里的历史时间没换算夏令时，导致非农、CPI 前后的挂单触发点和实盘差出一个小时。CDaylightSavings_US 这个类就是专门把美国每年的时钟切换点塞进数据库，让回测时间和经纪商服务器时间对齐。 它公开继承自 CDaylightSavings，构造上只暴露一个空参构造函数，真正的活儿在 SetDaylightSavings_US 里干。函数内部先 new 一个 CArrayObj 容器，再逐年 Add 具体的起止时刻。 代码里给出的数据从 2007 覆盖到 2017，每年开始都在 3 月第二个周日 03:00 拨快（如 2007.03.11、2016.03.13），结束在 11 月第一个周日 01:00 拨回（如 2007.11.04、2015.11.01）。打开 MT5 把这段 include 进你的 mqh，跑 2010 年 11 月那周，能直接验证 11.07 01:00 的回拨是否生效。 外汇和贵金属受数据行情冲击大、杠杆高，这类时间校准只解决回测偏差，不预示任何方向，实盘仍需自行核对经纪商时钟。
class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| NewsTrading | class=class="str">"cmt">//| Copyright class="num">2024, MetaQuotes Ltd. | class=class="str">"cmt">//| [MQL5官方文档] | class=class="str">"cmt">//+------------------------------------------------------------------+ class="macro">#include "DaylightSavings.mqh" class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|DaylightSavings_US Class | class=class="str">"cmt">//+------------------------------------------------------------------+ class CDaylightSavings_US: class="kw">public CDaylightSavings { class="kw">public: CDaylightSavings_US(class="type">void); }; class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Set Daylight Savings Schedule for the United States | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">void CDaylightSavings::SetDaylightSavings_US() { savings = new CArrayObj(); class=class="str">"cmt">//Daylight savings dates to readjust dates in the database for accurate testing in the strategy tester savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2007.03.class="num">11 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2007.11.class="num">04 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2008.03.class="num">09 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2008.11.class="num">02 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2009.03.class="num">08 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2009.11.class="num">01 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2010.03.class="num">14 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2010.11.class="num">07 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2011.03.class="num">13 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2011.11.class="num">06 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2012.03.class="num">11 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2012.11.class="num">04 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2013.03.class="num">10 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2013.11.class="num">03 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2014.03.class="num">09 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2014.11.class="num">02 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2015.03.class="num">08 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2015.11.class="num">01 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2016.03.class="num">13 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2016.11.class="num">06 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2017.03.class="num">12 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2017.11.class="num">05 class="num">01:class="num">00:class="num">00&class="macro">#x27;));
◍ 美式夏令时切换的硬编码清单
这段代码把 2018–2029 共 12 年的美国夏令时起止点，以 CDaylightSavings 对象逐条塞进 savings 列表。起始统一是当年 3 月第二个周日 03:00:00（例如 2018.03.11、2025.03.09），结束统一是当年 11 月第一个周日 01:00:00（例如 2018.11.04、2027.11.07），和美股盘面时钟跳变规则一致。 构造函数 CDaylightSavings_US 只做一件事：调 SetDaylightSavings_US() 把上面那组日期灌进实例。你若做跨时区 EA，直接抄这段能省掉手算夏令时的麻烦，但注意它写死了 2029 年之后就断了，2030 年需自己补行。 外汇与贵金属受夏令时跳点影响，点差和流动性可能在 03:00 EST 切换瞬间异常，属高风险时段，回测和实盘都要核对时区偏移。
savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2018.03.class="num">11 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2018.11.class="num">04 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2019.03.class="num">10 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2019.11.class="num">03 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2020.03.class="num">08 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2020.11.class="num">01 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2021.03.class="num">14 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2021.11.class="num">07 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2022.03.class="num">13 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2022.11.class="num">06 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2023.03.class="num">12 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2023.11.class="num">05 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2024.03.class="num">10 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2024.11.class="num">03 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2025.03.class="num">09 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2025.11.class="num">02 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2026.03.class="num">08 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2026.11.class="num">01 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2027.03.class="num">14 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2027.11.class="num">07 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2028.03.class="num">12 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2028.11.class="num">05 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); savings.Add(new CDaylightSavings(D&class="macro">#x27;class="num">2029.03.class="num">11 class="num">03:class="num">00:class="num">00&class="macro">#x27;,D&class="macro">#x27;class="num">2029.11.class="num">04 class="num">01:class="num">00:class="num">00&class="macro">#x27;)); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Constructor | class=class="str">"cmt">//+------------------------------------------------------------------+ CDaylightSavings_US::CDaylightSavings_US(class="type">void) { SetDaylightSavings_US(); } class=class="str">"cmt">//+------------------------------------------------------------------+
把交易品种属性收进一个类里
写 EA 时反复调用 SymbolInfoDouble、SymbolInfoInteger 去取点差、步长、冻结水平，代码会迅速变脏。CSymbolProperties 的思路是：在构造函数里把常用属性一次性缓存成成员变量，别的类直接读 ASK、BID、LOTSSTEP 这种字段，省掉到处查接口的红冗余。 构造函数里 ASK 先赋 0.0，是因为此时还没真正拿到报价；后面靠一个带可选 SYMBOL 参数的方法去填。若传 NULL，就默认取当前图表品种 Symbol()，找不到名就 Print 报错并返回 0.0，逻辑上把“取不到”和“取到 0”绑定，调用方得自己判。 属性清单里几个直接影响风控的：合约规模越大个人风险越高；最小/最大交易量卡住单仓手数，但多仓总和可能突破最大值；步长若为 1 就别想下 1.5 手；点差本质是经纪商抽的“税”，点差过高策略可能直接不盈利。外汇和贵金属波动大、杠杆高，这些参数读错就可能爆仓，务必在 MT5 里逐项核对。 标准化价格函数很实用：EURUSD 买价 1.07735 是 5 位精度，你传 1.077351 会报“无效价格”，它把第 6 位舍去变回 5 位。背景色那个 allow_black 默认 false，是因为后面拿品种色当图表底色，黑色会和 K 线、指标撞色没法看。
class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| NewsTrading | class=class="str">"cmt">//| Copyright class="num">2024, MetaQuotes Ltd. | class=class="str">"cmt">//| [MQL5官方文档] | class=class="str">"cmt">//+------------------------------------------------------------------+ class="macro">#include <Trade/SymbolInfo.mqh> class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|SymbolProperties class | class=class="str">"cmt">//+------------------------------------------------------------------+ class CSymbolProperties { class="kw">private: class="type">class="kw">double ASK;class=class="str">"cmt">//Store Ask Price class="type">class="kw">double BID;class=class="str">"cmt">//Store Bid Price class="type">class="kw">double LOTSMIN;class=class="str">"cmt">//Store Minimum Lotsize class="type">class="kw">double LOTSMAX;class=class="str">"cmt">//Store Maximum Lotsize class="type">class="kw">double LOTSSTEP;class=class="str">"cmt">//Store Lotsize Step class="type">class="kw">double LOTSLIMIT;class=class="str">"cmt">//Store Lotsize Limit(Maximum sum of Volume) class="type">long SPREAD;class=class="str">"cmt">//Store Spread value class="type">long STOPLEVEL;class=class="str">"cmt">//Store Stop level class="type">long FREEZELEVEL;class=class="str">"cmt">//Store Freeze level class="type">long TIME;class=class="str">"cmt">//Store time class="type">long DIGITS;class=class="str">"cmt">//Store Digits class="type">class="kw">double POINT;class=class="str">"cmt">//Store Point class="type">class="kw">double ORDERSVOLUME;class=class="str">"cmt">//Store Orders volume
「封装品种属性的类骨架」
在 MT5 里直接调 SymbolInfoDouble 之类函数虽然能跑，但每次传品种名、判返回值很啰嗦。把常用属性收进一个 CSymbolProperties 类，后面写 EA 或指标会清爽很多。 下面这段声明给出了类的核心成员：用 double 存持仓量 POSITIONSVOLUME，用 long 标记是否自定义品种 CUSTOM 和背景色 BACKGROUND_CLR；protected 区里直接挂了一个 CSymbolInfo 对象 CSymbol，所有底层查询都委托给它。 SetSymbolName 是内部设品种名的入口，若传 NULL 就默认取当前图表 Symbol()，设名失败会在日志打 Invalid Symbol 并返回 false。GetSymbolName 则原样吐回 CSymbol.Name()，对外隐藏了对象细节。 public 接口几乎覆盖了盯盘要用的所有字段：Ask/Bid 取报价，ContractSize 取合约规模，LotsMin/LotsMax/LotsStep/LotsLimit 管手数边界，Spread/StopLevel/FreezeLevel 是风控距离，Time 取品种时间，NormalizePrice 和 Digits 负责按点值规整价格。开 MT5 新建个类把这些声明贴进去，编译过即说明你的交易品种封装层立住了。
class="type">class="kw">double POSITIONSVOLUME;class=class="str">"cmt">//Store Positions volume class="type">long CUSTOM;class=class="str">"cmt">//Store if Symbol is Custom class="type">long BACKGROUND_CLR;class=class="str">"cmt">//Store Symbol&class="macro">#x27;s background class="type">class="kw">color class="kw">protected: CSymbolInfo CSymbol;class=class="str">"cmt">//Creating class CSymbolInfo&class="macro">#x27;s Object class="type">bool SetSymbolName(class="type">class="kw">string SYMBOL) { class=class="str">"cmt">//-- If Symbol&class="macro">#x27;s name was successfully set. if(!CSymbol.Name((SYMBOL==NULL)?Symbol():SYMBOL)) { Print("Invalid Symbol: ",SYMBOL); class="kw">return false; } class="kw">return true; } class=class="str">"cmt">//-- Retrieve Symbol&class="macro">#x27;s name class="type">class="kw">string GetSymbolName() { class="kw">return CSymbol.Name(); } class="kw">public: CSymbolProperties(class="type">void);class=class="str">"cmt">//Constructor class="type">class="kw">double Ask(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Ask Price class="type">class="kw">double Bid(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Bid Price class="type">class="kw">double ContractSize(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Contract Size class="type">class="kw">double LotsMin(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Min Volume class="type">class="kw">double LotsMax(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Max Volume class="type">class="kw">double LotsStep(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Volume Step class="type">class="kw">double LotsLimit(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Volume Limit class="type">int Spread(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Spread class="type">int StopLevel(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Stop Level class="type">int FreezeLevel(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Freeze Level class="type">class="kw">datetime Time(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Symbol&class="macro">#x27;s Time class=class="str">"cmt">//-- Normalize Price class="type">class="kw">double NormalizePrice(const class="type">class="kw">double price,class="type">class="kw">string SYMBOL=NULL); class="type">int Digits(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Symbol&class="macro">#x27;s Digits
◍ 封装品种属性读取的 C++ 类接口
做 MT5 ea 或指标时，反复调用 SymbolInfoDouble 会让主逻辑变脏。把常用品种属性收进一个 CSymbolProperties 类，用成员函数统一取数，是更干净的做法。 下面这组公开方法覆盖了点值、交易模式、挂单与持仓总仓、基础/利润/保证金币种等。注意每个函数都允许传 SYMBOL 参数，缺省为 NULL 即当前图表品种： double Point(string SYMBOL=NULL); // 取品种点值 ENUM_SYMBOL_TRADE_MODE TradeMode(string SYMBOL=NULL); // 取交易模式 double OrdersVolume(string SYMBOL=NULL); // 取挂单总仓 double PositionsVolume(string SYMBOL=NULL); // 取持仓总仓 string CurrencyBase(string SYMBOL=NULL); // 基础币种 string CurrencyProfit(string SYMBOL=NULL); // 利润币种 string CurrencyMargin(string SYMBOL=NULL); // 保证金币种 bool Custom(string SYMBOL=NULL); // 是否自定义品种 color SymbolBackground(string SYMBOL=NULL,bool allow_black=false); // 背景色 构造函数把所有缓存字段清零：ASK、BID 为 0.0，DIGITS 为 0，SPREAD、STOPLEVEL、FREEZELEVEL、TIME、CUSTOM、BACKGROUND_CLR 均为 0 或 0.0。这样对象刚 new 出来时不会带脏数据。 以 Ask 为例，内部先 SetSymbolName 切换品种，再走 CSymbol.InfoDouble(SYMBOL_ASK,ASK)；失败则 Print 报错并返回 0.0。Bid 逻辑完全一致，只是换 SYMBOL_BID。外汇与贵金属杠杆高，回测时若 SYMBOL 名写错导致返回 0.0，可能让仓位计算直接崩坏，实盘前务必在策略测试器里打印验证。
class="type">class="kw">double Point(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Symbol&class="macro">#x27;s Point ENUM_SYMBOL_TRADE_MODE TradeMode(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Symbol&class="macro">#x27;s Trade Mode class="type">class="kw">double OrdersVolume(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Symbol&class="macro">#x27;s Orders Volume class="type">class="kw">double PositionsVolume(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Symbol&class="macro">#x27;s Positions Volume class="type">class="kw">string CurrencyBase(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Symbol&class="macro">#x27;s Currency Base class="type">class="kw">string CurrencyProfit(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Symbol&class="macro">#x27;s Currency Profit class="type">class="kw">string CurrencyMargin(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Symbol&class="macro">#x27;s Currency Margin class="type">bool Custom(class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Symbol&class="macro">#x27;s Custom status class="type">class="kw">color SymbolBackground(class="type">class="kw">string SYMBOL=NULL,class="type">bool allow_black=false);class=class="str">"cmt">//Retrieve Symbol&class="macro">#x27;s Background class="type">class="kw">color }; class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Constructor | class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//Initializing Variables CSymbolProperties::CSymbolProperties(class="type">void):ASK(class="num">0.0),BID(class="num">0.0), LOTSMIN(class="num">0.0),LOTSMAX(class="num">0.0), LOTSSTEP(class="num">0.0),LOTSLIMIT(class="num">0.0),DIGITS(class="num">0), SPREAD(class="num">0),STOPLEVEL(class="num">0),ORDERSVOLUME(class="num">0.0), FREEZELEVEL(class="num">0),TIME(class="num">0),POINT(class="num">0.0),POSITIONSVOLUME(class="num">0.0), CUSTOM(class="num">0),BACKGROUND_CLR(class="num">0) { } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Ask Price | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">double CSymbolProperties::Ask(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_ASK,ASK)) { class="kw">return ASK; } } Print("Unable to retrieve Symbol&class="macro">#x27;s Ask Price"); class="kw">return class="num">0.0; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Bid Price | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">double CSymbolProperties::Bid(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_BID,BID)) { class="kw">return BID; } } Print("Unable to retrieve Symbol&class="macro">#x27;s Bid Price");
把合约规格和手数限制一次性读出来
做贵金属或外汇跨品种下单前，最容易被忽略的是各品种合约大小和手数边界不一致。XAUUSD 的合约大小通常是 100 盎司，EURUSD 多为 100000 单位，若硬编码容易在调仓时算错名义价值。 下面这组方法封装了从交易品种对象取属性的逻辑：ContractSize 走 Refresh 后取合约大小，LotsMin / LotsMax / LotsStep / LotsLimit 则直接调 InfoDouble 拿对应枚举值。任一环节失败都会 Print 报错并返回 0.0，调用方必须判断非零才可信。 实盘里建议对监控列表每个符号启动时跑一遍，把返回值和券商展示的最小 0.01、最大 50 手等做交叉核对，能提前发现少数平台 LotsLimit 返回 0 表示无限制这一坑。外汇与贵金属杠杆高，手数越界会直接拒单，属于高频但可预防的失误。
class="type">class="kw">double CSymbolProperties::ContractSize(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.Refresh()) { class="kw">return CSymbol.ContractSize(); } } Print("Unable to retrieve Symbol&class="macro">#x27;s Contract size"); class="kw">return class="num">0.0; } class="type">class="kw">double CSymbolProperties::LotsMin(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_VOLUME_MIN,LOTSMIN)) { class="kw">return LOTSMIN; } } Print("Unable to retrieve Symbol&class="macro">#x27;s LotsMin"); class="kw">return class="num">0.0; } class="type">class="kw">double CSymbolProperties::LotsMax(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_VOLUME_MAX,LOTSMAX)) { class="kw">return LOTSMAX; } } Print("Unable to retrieve Symbol&class="macro">#x27;s LotsMax"); class="kw">return class="num">0.0; } class="type">class="kw">double CSymbolProperties::LotsStep(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_VOLUME_STEP,LOTSSTEP)) { class="kw">return LOTSSTEP; } } Print("Unable to retrieve Symbol&class="macro">#x27;s LotsStep"); class="kw">return class="num">0.0; } class="type">class="kw">double CSymbolProperties::LotsLimit(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_VOLUME_LIMIT,LOTSLIMIT)) { class="kw">return LOTSLIMIT; } }
「抓点差与止损冻结位的类方法」
在封装品种属性的类里，点差、止损位、冻结位这些交易约束值都走同一套取数逻辑：先切到目标品种，再用 InfoInteger 读对应宏，失败就打印错误并返回 0。 Spread 方法读 SYMBOL_SPREAD，返回 int 类型的点差值；StopLevel 读 SYMBOL_TRADE_STOPS_LEVEL，FreezeLevel 读 SYMBOL_TRADE_FREEZE_LEVEL，三者失败均返回 0，意味着调用方拿到 0 时要警惕可能是取数失败而非真为 0。 Time 方法取 SYMBOL_TIME，失败时显式把 TIME 置 0 再返回 datetime(0)，和外部返回 0.0 的 LotsLimit 风格略有差异，但本质都是用返回值兜底。 下面这段是 Spread 到 Time 的紧凑实现，注意每个方法都先调 SetSymbolName 再读属性：
class="type">int CSymbolProperties::Spread(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_SPREAD,SPREAD)) { class="kw">return class="type">int(SPREAD); } } Print("Unable to retrieve Symbol&class="macro">#x27;s Spread"); class="kw">return class="num">0; } class="type">int CSymbolProperties::StopLevel(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_TRADE_STOPS_LEVEL,STOPLEVEL)) { class="kw">return class="type">int(STOPLEVEL); } } Print("Unable to retrieve Symbol&class="macro">#x27;s StopLevel"); class="kw">return class="num">0; } class="type">int CSymbolProperties::FreezeLevel(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_TRADE_FREEZE_LEVEL,FREEZELEVEL)) { class="kw">return class="type">int(FREEZELEVEL); } } Print("Unable to retrieve Symbol&class="macro">#x27;s FreezeLevel"); class="kw">return class="num">0; } class="type">class="kw">datetime CSymbolProperties::Time(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_TIME,TIME)) { class="kw">return class="type">class="kw">datetime(TIME); } } Print("Unable to retrieve Symbol&class="macro">#x27;s Time"); TIME=class="num">0; class="kw">return class="type">class="kw">datetime(TIME); }
◍ 封装品种属性读取的容错逻辑
在 MT5 自建指标或 EA 时，直接调 Terminal 信息函数经常拿到脏值，尤其是切换品种或报价刷新滞后的时候。下面这段封装把价格归一化、小数位、点值、交易模式、挂单总量都包进 CSymbolProperties 类方法里，先 Refresh 再取数，取不到就打日志返回安全默认值。 价格归一化那段，Refresh 和 RefreshRates 都成功才走 NormalizePrice，否则原样返回 price 并提示失败——外汇和贵金属点差跳变时这层保护能避免下错tick尺寸的伪信号，相关品种波动高风险需自担。 Digits 和 Point 分别用 InfoInteger(SYMBOL_DIGITS) 和 InfoDouble(SYMBOL_POINT) 拿，失败返回 0 和 0.0；XAUUSD 正常是 2 位小数、point=0.01，若返回 0 说明品种上下文没挂上，得查 SetSymbolName 调用。 TradeMode 调 Refresh 后取枚举，异常返回 SYMBOL_TRADE_MODE_DISABLED；OrdersVolume 则遍历 OrdersTotal 用 OrderSelect+OrderGetString 比对符号累加，这部分循环在挂单多时可能偏慢，实盘建议加缓存。
if(CSymbol.Refresh()&&CSymbol.RefreshRates()) { class="kw">return CSymbol.NormalizePrice(price); } } Print("Unable to Normalize Symbol&class="macro">#x27;s Price"); class="kw">return price; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Symbol&class="macro">#x27;s Digits | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">int CSymbolProperties::Digits(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_DIGITS,DIGITS)) { class="kw">return class="type">int(DIGITS); } } Print("Unable to retrieve Symbol&class="macro">#x27;s Digits"); class="kw">return class="num">0; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Symbol&class="macro">#x27;s Point | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">double CSymbolProperties::Point(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_POINT,POINT)) { class="kw">return POINT; } } Print("Unable to retrieve Symbol&class="macro">#x27;s Point"); class="kw">return class="num">0.0; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Symbol&class="macro">#x27;s Trade Mode | class=class="str">"cmt">//+------------------------------------------------------------------+ ENUM_SYMBOL_TRADE_MODE CSymbolProperties::TradeMode(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.Refresh()) { class="kw">return CSymbol.TradeMode(); } } Print("Unable to retrieve Symbol&class="macro">#x27;s TradeMode"); class="kw">return SYMBOL_TRADE_MODE_DISABLED; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Symbol&class="macro">#x27;s Orders Volume | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">double CSymbolProperties::OrdersVolume(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { for(class="type">int i=class="num">0; i<OrdersTotal(); i++) { if(OrderSelect(OrderGetTicket(i))) { if(OrderGetString(ORDER_SYMBOL)==GetSymbolName()) {
挂单与持仓量的类内抓取逻辑
在封装品种属性的类里，挂单总量和持仓总量是分开算的。OrdersVolume 方法遍历 OrdersTotal()，只累加当前品种且 ticket 有效的 ORDER_VOLUME_CURRENT；PositionsVolume 同理遍历 PositionsTotal()，用 POSITION_SYMBOL 比对本地品种名再累加 POSITION_VOLUME。 两个方法都依赖 SetSymbolName 先切换上下文，失败就 Print 报错并返回 0.0。这意味着若品种名设错，返回值是 0 而不是 NaN，回测时容易误判为「无仓位」。 CurrencyBase / CurrencyProfit 则走另一条路：先 CSymbol.Refresh() 再调 CSymbol 的对应方法，失败返回空串。注意 Refresh 不成功时连 Print 都不进 if 分支，直接落到函数末尾报错。 开 MT5 把这段塞进 EA 的 OnTick 前面，打印 EURUSD 的 OrdersVolume 与 PositionsVolume，能直接看到当前品种挂单和持仓的裸数值，验证类方法是否和你账户实际订单对齐。
ORDERSVOLUME+=OrderGetDouble(ORDER_VOLUME_CURRENT); } } } } else { Print("Unable to retrieve Symbol&class="macro">#x27;s OrdersVolume"); class="kw">return class="num">0.0; } class="kw">return ORDERSVOLUME; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Symbol&class="macro">#x27;s Positions Volume | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">double CSymbolProperties::PositionsVolume(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { for(class="type">int i=class="num">0; i<PositionsTotal(); i++) { if(PositionGetTicket(i)>class="num">0) { if(PositionGetString(POSITION_SYMBOL)==GetSymbolName()) { POSITIONSVOLUME+=PositionGetDouble(POSITION_VOLUME); } } } } else { Print("Unable to retrieve Symbol&class="macro">#x27;s PositionsVolume"); class="kw">return class="num">0.0; } class="kw">return POSITIONSVOLUME; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Symbol&class="macro">#x27;s Currency Base | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">string CSymbolProperties::CurrencyBase(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.Refresh()) { class="kw">return CSymbol.CurrencyBase(); } } Print("Unable to retrieve Symbol&class="macro">#x27;s CurrencyBase"); class="kw">return ""; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Symbol&class="macro">#x27;s Currency Profit | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">string CSymbolProperties::CurrencyProfit(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.Refresh()) { class="kw">return CSymbol.CurrencyProfit(); } } Print("Unable to retrieve Symbol&class="macro">#x27;s CurrencyProfit"); class="kw">return ""; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Symbol&class="macro">#x27;s Currency Margin | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">string CSymbolProperties::CurrencyMargin(class="type">class="kw">string SYMBOL=NULL) {
「读取品种保证金与自定义状态的取值细节」
在封装品种属性的类里，取保证金和判自定义这两件事都依赖先切到目标品种。切品种用 SetSymbolName，失败就直接 Print 报错并返回空值或 false，不会让后续计算拿到脏数据。 CurrencyMargin 的取法是先 Refresh 再返回数值；Custom 则走 InfoInteger(SYMBOL_CUSTOM) 把结果转成 bool。两者都带兜底：取不到就打印 'Unable to retrieve…' 并返回安全默认值。 背景色那段有个坑：默认不允许黑色背景，若取到 '0,0,0' 或 clrBlack 且 allow_black=false，会替换成浅灰 '236,236,236'。这个 236,236,236 是 RGB 十进制，对应 MT5 里接近 C'236,236,236' 的浅灰，避免黑底白字在面板上看不清。 构造函数把所有成员初始化为零值：ASK/BID/LOTSMIN 等是 0.0，DIGITS/SPREAD/STOPLEVEL 是 0。开 MT5 把这段贴进 CSymbolProperties 类，编译后调用 Custom(_Symbol) 能直观看到当前品种是否自定义品种（返回 true 概率出现在券商自建符号上）。外汇与贵金属杠杆品种波动剧烈，这类属性读取仅用于风控展示，不预示价格方向。
if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.Refresh()) { class="kw">return CSymbol.CurrencyMargin(); } } Print("Unable to retrieve Symbol&class="macro">#x27;s CurrencyMargin"); class="kw">return ""; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Symbol&class="macro">#x27;s Custom status | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">bool CSymbolProperties::Custom(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_CUSTOM,CUSTOM)) { class="kw">return class="type">bool(CUSTOM); } } Print("Unable to retrieve if Symbol is Custom"); class="kw">return false; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Symbol&class="macro">#x27;s Background class="type">class="kw">color | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">color CSymbolProperties::SymbolBackground(class="type">class="kw">string SYMBOL=NULL,class="type">bool allow_black=false) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_BACKGROUND_COLOR,BACKGROUND_CLR)) { class=class="str">"cmt">/*Avoid any Symbol black background class="type">class="kw">color */ BACKGROUND_CLR = ((ColorToString(class="type">class="kw">color(BACKGROUND_CLR))=="class="num">0,class="num">0,class="num">0"|| class="type">class="kw">color(BACKGROUND_CLR)==clrBlack)&&!allow_black)? class="type">long(StringToColor("class="num">236,class="num">236,class="num">236")):BACKGROUND_CLR; class="kw">return class="type">class="kw">color(BACKGROUND_CLR); } } Print("Unable to retrieve Symbol&class="macro">#x27;s Background class="type">class="kw">color"); class="kw">return class="type">class="kw">color(StringToColor("class="num">236,class="num">236,class="num">236"));class=class="str">"cmt">//Retrieve a lightish gray class="type">class="kw">color } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//Initializing Variables CSymbolProperties::CSymbolProperties(class="type">void):ASK(class="num">0.0),BID(class="num">0.0), LOTSMIN(class="num">0.0),LOTSMAX(class="num">0.0), LOTSSTEP(class="num">0.0),LOTSLIMIT(class="num">0.0),DIGITS(class="num">0), SPREAD(class="num">0),STOPLEVEL(class="num">0),ORDERSVOLUME(class="num">0.0), FREEZELEVEL(class="num">0),TIME(class="num">0),POINT(class="num">0.0),POSITIONSVOLUME(class="num">0.0), CUSTOM(class="num">0),BACKGROUND_CLR(class="num">0) { } class="type">class="kw">double CSymbolProperties::Ask(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL)) {
◍ 把品种属性封装成可复用取数函数
在 MT5 自定义类里，把 Ask、Bid、合约大小、最小/最大/步进手数全部做成带默认参数的方法，能省掉反复写 SymbolInfoDouble 的麻烦。下面这组实现统一走 SetSymbolName 设置品种，再调底层接口取数，失败时打印并返回 0.0。 以 Bid 方法为例：传入 SYMBOL 为空时自动用当前图表的 Symbol()，先 SetSymbolName 校验品种名，再 CSymbol.InfoDouble(SYMBOL_BID,BID) 取买价。若返回 false 或取数失败，终端会输出 "Unable to retrieve Symbol's Bid Price"，调用方拿到 0.0 要自行判断。 ContractSize 多一步 CSymbol.Refresh()，确保每次取合约规模前刷新品种缓存，避免跨品种切换时读到旧值。LotsMin / LotsMax / LotsStep 分别对应 SYMBOL_VOLUME_MIN、SYMBOL_VOLUME_MAX、SYMBOL_VOLUME_STEP，外汇品种常见 MIN=0.01、STEP=0.01，但黄金 XAUUSD 的 MAX 可能到 1000 以上，硬编常数会出错。 开 MT5 建个 mqh 把这段代码贴进类里，在 EA 里用 CSymbolProperties().LotsStep("XAUUSD") 跑一下，就能直接拿到该品种真实步进手数，比每次手写 SymbolInfoDouble 稳。
class="type">class="kw">double CSymbolProperties::Bid(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_BID,BID)) { class="kw">return BID; } } Print("Unable to retrieve Symbol&class="macro">#x27;s Bid Price"); class="kw">return class="num">0.0; } class="type">class="kw">double CSymbolProperties::ContractSize(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.Refresh()) { class="kw">return CSymbol.ContractSize(); } } Print("Unable to retrieve Symbol&class="macro">#x27;s Contract size"); class="kw">return class="num">0.0; } class="type">class="kw">double CSymbolProperties::LotsMin(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_VOLUME_MIN,LOTSMIN)) { class="kw">return LOTSMIN; } } Print("Unable to retrieve Symbol&class="macro">#x27;s LotsMin"); class="kw">return class="num">0.0; } class="type">class="kw">double CSymbolProperties::LotsMax(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_VOLUME_MAX,LOTSMAX)) { class="kw">return LOTSMAX; } } Print("Unable to retrieve Symbol&class="macro">#x27;s LotsMax"); class="kw">return class="num">0.0; } class="type">class="kw">double CSymbolProperties::LotsStep(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_VOLUME_STEP,LOTSSTEP)) { class="kw">return LOTSSTEP; } } Print("Unable to retrieve Symbol&class="macro">#x27;s LotsStep"); class="kw">return class="num">0.0; }
封装品种交易限制与报价的取数接口
在写 EA 或指标时，直接反复调用 Terminal 自带的 SymbolInfo 系列函数会让代码散落各处。下面这组方法把常用品种属性收口到一个类里，传入品种名（缺省为 NULL 即当前图表品种）就能拿到数值，调用前统一走 SetSymbolName 做上下文切换。 LotsLimit 返回该品种单笔最大允许手数（SYMBOL_VOLUME_LIMIT），失败返回 0.0；Spread 取点差整数值（SYMBOL_SPREAD），StopLevel 与 FreezeLevel 分别取止损水平和冻结层级（SYMBOL_TRADE_STOPS_LEVEL / SYMBOL_TRADE_FREEZE_LEVEL），均失败返回 0。 Time 方法拿品种最后报价时间（SYMBOL_TIME），若取不到会把内部 TIME 显式置 0 再返回，这点和前面几个方法略有差异——前者只 Print 报错，这里多一步归零。NormalizePrice 先 Refresh 再 RefreshRates，成功才用 CSymbol.NormalizePrice 规整价格，否则原样返回，避免乱跳小数位导致下单被拒。 外汇与贵金属杠杆高、点差跳动快，这类封装能让你在回测或实盘里一键查到当前品种限制，开 MT5 把代码贴进 include 文件即可验证返回值。
class="type">class="kw">double CSymbolProperties::LotsLimit(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_VOLUME_LIMIT,LOTSLIMIT)) { class="kw">return LOTSLIMIT; } } Print("Unable to retrieve Symbol&class="macro">#x27;s LotsLimit"); class="kw">return class="num">0.0; } class="type">int CSymbolProperties::Spread(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_SPREAD,SPREAD)) { class="kw">return class="type">int(SPREAD); } } Print("Unable to retrieve Symbol&class="macro">#x27;s Spread"); class="kw">return class="num">0; } class="type">int CSymbolProperties::StopLevel(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_TRADE_STOPS_LEVEL,STOPLEVEL)) { class="kw">return class="type">int(STOPLEVEL); } } Print("Unable to retrieve Symbol&class="macro">#x27;s StopLevel"); class="kw">return class="num">0; } class="type">int CSymbolProperties::FreezeLevel(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_TRADE_FREEZE_LEVEL,FREEZELEVEL)) { class="kw">return class="type">int(FREEZELEVEL); } } Print("Unable to retrieve Symbol&class="macro">#x27;s FreezeLevel"); class="kw">return class="num">0; } class="type">class="kw">datetime CSymbolProperties::Time(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_TIME,TIME)) { class="kw">return class="type">class="kw">datetime(TIME); } } Print("Unable to retrieve Symbol&class="macro">#x27;s Time"); TIME=class="num">0; class="kw">return class="type">class="kw">datetime(TIME); } class="type">class="kw">double CSymbolProperties::NormalizePrice(const class="type">class="kw">double price,class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.Refresh()&&CSymbol.RefreshRates()) { class="kw">return CSymbol.NormalizePrice(price); } } Print("Unable to Normalize Symbol&class="macro">#x27;s Price"); class="kw">return price; } class="type">int CSymbolProperties::Digits(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_DIGITS,DIGITS)) {
「从品种属性类里抠出实时盘口数据」
在封装好的 CSymbolProperties 类里，几个成员函数专门负责把 MT5 当前品种的微观参数和挂单持仓体量捞出来。Digits() 和 Point() 都先走 SetSymbolName 切换目标品种，再调底层 Info 接口拿精度与最小报价单位；拿不到就 Print 报错并返回 0 或 0.0，EURUSD 通常返回 DIGITS=5、POINT=0.00001，XAUUSD 多为 DIGITS=2、POINT=0.01。 TradeMode() 多一步 CSymbol.Refresh() 强制刷新，避免读到缓存里的旧交易权限状态，返回枚举 SYMBOL_TRADE_MODE_DISABLED 代表该品种暂时禁交易，这种事在重要数据发布前后真会发生。 OrdersVolume() 与 PositionsVolume() 则是遍历全局订单池和持仓池，用 ORDER_SYMBOL / POSITION_SYMBOL 比对当前品种名做累加。ORDERSVOLUME 取 ORDER_VOLUME_CURRENT（挂单余量），POSITIONSVOLUME 取 POSITION_VOLUME（已开仓量）；若 SetSymbolName 失败直接返回 0.0，不会漏加但也不抛异常。 开 MT5 按 F4 把这段塞进自己的 EA 框架，接上 Print 输出，切到 XAUUSD 跑一遍，能看到贵金属在高杠杆下的持仓体积累速度明显快于直盘，外汇贵金属波动剧烈、杠杆风险高，参数验证后请自行评估仓位。
class="kw">return class="type">int(DIGITS); } } Print("Unable to retrieve Symbol&class="macro">#x27;s Digits"); class="kw">return class="num">0; } class="type">class="kw">double CSymbolProperties::Point(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoDouble(SYMBOL_POINT,POINT)) { class="kw">return POINT; } } Print("Unable to retrieve Symbol&class="macro">#x27;s Point"); class="kw">return class="num">0.0; } ENUM_SYMBOL_TRADE_MODE CSymbolProperties::TradeMode(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.Refresh()) { class="kw">return CSymbol.TradeMode(); } } Print("Unable to retrieve Symbol&class="macro">#x27;s TradeMode"); class="kw">return SYMBOL_TRADE_MODE_DISABLED; } class="type">class="kw">double CSymbolProperties::OrdersVolume(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { for(class="type">int i=class="num">0; i<OrdersTotal(); i++) { if(OrderSelect(OrderGetTicket(i))) { if(OrderGetString(ORDER_SYMBOL)==GetSymbolName()) { ORDERSVOLUME+=OrderGetDouble(ORDER_VOLUME_CURRENT); } } } } else { Print("Unable to retrieve Symbol&class="macro">#x27;s OrdersVolume"); class="kw">return class="num">0.0; } class="kw">return ORDERSVOLUME; } class="type">class="kw">double CSymbolProperties::PositionsVolume(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { for(class="type">int i=class="num">0; i<PositionsTotal(); i++) { if(PositionGetTicket(i)>class="num">0) { if(PositionGetString(POSITION_SYMBOL)==GetSymbolName()) { POSITIONSVOLUME+=PositionGetDouble(POSITION_VOLUME); } } } } else { Print("Unable to retrieve Symbol&class="macro">#x27;s PositionsVolume"); class="kw">return class="num">0.0; } class="kw">return POSITIONSVOLUME; }
◍ 从交易品种扒出币种与背景色
在 MT5 的自定义指标或 EA 里，经常要确认当前品种的基础币、盈利币、保证金币，以及它图表背景是不是纯黑。下面这组方法把 CSymbol 的封装调用拆得很直白，复制进你的类就能直接用。 CurrencyBase / CurrencyProfit / CurrencyMargin 三个函数结构完全一致：先 SetSymbolName 切换品种，再 CSymbol.Refresh() 拉一次最新属性，成功就返回对应字符串，失败 Print 报错并返回空串。实战中若返回空，大概率是你传的 SYMBOL 名字拼错或该品种未在市场报价窗口。 Custom 函数走的是 InfoInteger(SYMBOL_CUSTOM, CUSTOM) 路径，把返回的整型强转 bool，用来判断该品种是不是自定义符号（比如某些券商的现货黄金变种）。 SymbolBackground 有个细节：默认拒绝纯黑背景（ColorToString 等于 "0,0,0" 或 clrBlack 且 allow_black=false 时），强制替换成 "236,236,236" 浅灰，避免黑底指标线看不清；传 allow_black=true 才放行。外汇与贵金属杠杆高、点差跳变快，用这套前先在策略测试器用 XAUUSD 跑一遍确认背景与币种返回值符合预期。
class="type">class="kw">string CSymbolProperties::CurrencyBase(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.Refresh()) { class="kw">return CSymbol.CurrencyBase(); } } Print("Unable to retrieve Symbol&class="macro">#x27;s CurrencyBase"); class="kw">return ""; } class="type">class="kw">string CSymbolProperties::CurrencyProfit(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.Refresh()) { class="kw">return CSymbol.CurrencyProfit(); } } Print("Unable to retrieve Symbol&class="macro">#x27;s CurrencyProfit"); class="kw">return ""; } class="type">class="kw">string CSymbolProperties::CurrencyMargin(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.Refresh()) { class="kw">return CSymbol.CurrencyMargin(); } } Print("Unable to retrieve Symbol&class="macro">#x27;s CurrencyMargin"); class="kw">return ""; } class="type">bool CSymbolProperties::Custom(class="type">class="kw">string SYMBOL=NULL) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_CUSTOM,CUSTOM)) { class="kw">return class="type">bool(CUSTOM); } } Print("Unable to retrieve if Symbol is Custom"); class="kw">return false; } class="type">class="kw">color CSymbolProperties::SymbolBackground(class="type">class="kw">string SYMBOL=NULL,class="type">bool allow_black=false) { if(SetSymbolName(SYMBOL))class=class="str">"cmt">//Set Symbol { if(CSymbol.InfoInteger(SYMBOL_BACKGROUND_COLOR,BACKGROUND_CLR)) { class=class="str">"cmt">/*Avoid any Symbol black background class="type">class="kw">color */ BACKGROUND_CLR = ((ColorToString(class="type">class="kw">color(BACKGROUND_CLR))=="class="num">0,class="num">0,class="num">0"|| class="type">class="kw">color(BACKGROUND_CLR)==clrBlack)&&!allow_black)? class="type">long(StringToColor("class="num">236,class="num">236,class="num">236")):BACKGROUND_CLR; class="kw">return class="type">class="kw">color(BACKGROUND_CLR); } } Print("Unable to retrieve Symbol&class="macro">#x27;s Background class="type">class="kw">color"); class="kw">return class="type">class="kw">color(StringToColor("class="num">236,class="num">236,class="num">236"));class=class="str">"cmt">//Retrieve a lightish gray class="type">class="kw">color }
时间操作类的接口设计
做新闻交易或时段过滤时，反复手写时间比对很拖累开发效率。把常用的时间换算与区间判断收进一个管理类，能让 EA 逻辑清爽不少。 下面这个类声明给出了可直接抄进 MT5 头文件的接口骨架：两个私有 MqlDateTime 变量存当天与格式对象；公开方法覆盖区间判定（DateIsInRange 有两个重载，分别判断单点与区间是否落在参照范围内）、今日判定，以及把各时间单位换算成秒的辅助函数（SecondsS 到 YearsS，都带默认乘数为 1 的 multiple 参数）。 实际验证时，在 MT5 里新建 EA 附带该类，调用 MinutesS(15) 应返回 900，HoursS(2) 返回 7200；外汇与贵金属品种受数据跳空影响，时间窗口策略回测须标注高风险，历史吻合不代表未来概率稳定。 别把类当黑盒 ReturnYear / ReturnMonth / ReturnDay 只是对 datetime 做结构体拆解，若传入 1970 年前的非法值，MT5 可能返回 0 或乱码，写策略前先打印自检。
class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| NewsTrading | class=class="str">"cmt">//| Copyright class="num">2024, MetaQuotes Ltd. | class=class="str">"cmt">//| [MQL5官方文档] | class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|TimeManagement class | class=class="str">"cmt">//+------------------------------------------------------------------+ class CTimeManagement { class="kw">private: class="type">MqlDateTime today;class=class="str">"cmt">//class="kw">private variable class="type">MqlDateTime timeFormat;class=class="str">"cmt">//class="kw">private variable class="kw">public: class=class="str">"cmt">//-- Checks if a date is within two other dates class="type">bool DateIsInRange(class="type">class="kw">datetime FirstTime,class="type">class="kw">datetime SecondTime,class="type">class="kw">datetime compareTime); class=class="str">"cmt">//-- Check if two dates(Start&End) are within CompareStart & CompareEnd class="type">bool DateIsInRange(class="type">class="kw">datetime Start,class="type">class="kw">datetime End,class="type">class="kw">datetime CompareStart,class="type">class="kw">datetime CompareEnd); class="type">bool DateisToday(class="type">class="kw">datetime TimeRepresented);class=class="str">"cmt">//Checks if a date is within the current day class="type">int SecondsS(class="type">int multiple=class="num">1);class=class="str">"cmt">//Returns seconds class="type">int MinutesS(class="type">int multiple=class="num">1);class=class="str">"cmt">//Returns Minutes in seconds class="type">int HoursS(class="type">int multiple=class="num">1);class=class="str">"cmt">//Returns Hours in seconds class="type">int DaysS(class="type">int multiple=class="num">1);class=class="str">"cmt">//Returns Days in seconds class="type">int WeeksS(class="type">int multiple=class="num">1);class=class="str">"cmt">//Returns Weeks in seconds class="type">int MonthsS(class="type">int multiple=class="num">1);class=class="str">"cmt">//Returns Months in seconds class="type">int YearsS(class="type">int multiple=class="num">1);class=class="str">"cmt">//Returns Years in seconds class="type">int ReturnYear(class="type">class="kw">datetime time);class=class="str">"cmt">//Returns the Year for a specific date class="type">int ReturnMonth(class="type">class="kw">datetime time);class=class="str">"cmt">//Returns the Month for a specific date class="type">int ReturnDay(class="type">class="kw">datetime time);class=class="str">"cmt">//Returns the Day for a specific date class=class="str">"cmt">//-- Will class="kw">return a class="type">class="kw">datetime type of a date with an subtraction offset in seconds
「时间类的秒数换算与区间判定」
CTimeManagement 里这组方法把『时间比较』和『周期换算』拆得很碎，但实盘里很好用。先做区间判定：单点是否在两段之间用 DateIsInRange(FirstTime, SecondTime, compareTime)，返回 FirstTime<=compareTime && SecondTime>compareTime；四点重载则判断 [Start,End] 是否整体落在 [CompareStart,CompareEnd] 内，条件是 Start<=CompareStart && CompareEnd<End。 DateisToday() 靠 TimeToStruct 拆出年月日，再和 TimeCurrent() 取的当天结构比对，三者全等才认作今天。注意它用的是服务器时间，跨时区策略要自己补偏移。 秒数换算链是嵌套乘法：SecondsS 返回 1*multiple，MinutesS 调 SecondsS(60)，HoursS 调 MinutesS(60)，DaysS 调 HoursS(24)，WeeksS 调 DaysS(7)，MonthsS 调 WeeksS(4)。也就是说 MonthsS(1) 实际是 60*60*24*7*4 = 2,419,200 秒——这是个近似月，不校闰月和天数差，回测长周期时得心里有数。 外汇和贵金属杠杆高、跳空频繁，用这类时间窗过滤信号时，先开 MT5 把 DateIsInRange 接进 EA 的 OnTick 跑一周模拟，确认边界开闭符合你的触发逻辑再上真仓。
class="type">class="kw">datetime TimeMinusOffset(class="type">class="kw">datetime standardtime,class="type">int timeoffset); class=class="str">"cmt">//-- Will class="kw">return a class="type">class="kw">datetime type of a date with an addition offset in seconds class="type">class="kw">datetime TimePlusOffset(class="type">class="kw">datetime standardtime,class="type">int timeoffset); }; class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Checks if a date is within two other dates | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">bool CTimeManagement::DateIsInRange(class="type">class="kw">datetime FirstTime,class="type">class="kw">datetime SecondTime,class="type">class="kw">datetime compareTime) { class="kw">return(FirstTime<=compareTime&&SecondTime>compareTime); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Check if two dates(Start&End) are within CompareStart & CompareEnd| class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">bool CTimeManagement::DateIsInRange(class="type">class="kw">datetime Start,class="type">class="kw">datetime End,class="type">class="kw">datetime CompareStart,class="type">class="kw">datetime CompareEnd) { class="kw">return(Start<=CompareStart&&CompareEnd<End); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Checks if a date is within the current day | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">bool CTimeManagement::DateisToday(class="type">class="kw">datetime TimeRepresented) { class="type">MqlDateTime TiM; TimeToStruct(TimeRepresented,TiM); TimeCurrent(today); class="kw">return(TiM.year==today.year&&TiM.mon==today.mon&&TiM.day==today.day); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Returns seconds | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">int CTimeManagement::SecondsS(class="type">int multiple=class="num">1) { class="kw">return (class="num">1*multiple); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Returns Minutes in seconds | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">int CTimeManagement::MinutesS(class="type">int multiple=class="num">1) { class="kw">return (SecondsS(class="num">60)*multiple); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Returns Hours in seconds | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">int CTimeManagement::HoursS(class="type">int multiple=class="num">1) { class="kw">return (MinutesS(class="num">60)*multiple); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Returns Days in seconds | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">int CTimeManagement::DaysS(class="type">int multiple=class="num">1) { class="kw">return (HoursS(class="num">24)*multiple); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Returns Weeks in seconds | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">int CTimeManagement::WeeksS(class="type">int multiple=class="num">1) { class="kw">return (DaysS(class="num">7)*multiple); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Returns Months in seconds | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">int CTimeManagement::MonthsS(class="type">int multiple=class="num">1) { class="kw">return (WeeksS(class="num">4)*multiple); }
◍ 时间管理类的年月日与偏移拆解
在 MT5 自研 EA 里，把时间换算封装成一个 CTimeManagement 类能少写很多重复代码。下面这组方法覆盖了「年秒数推算」和「从 datetime 拆年月日」两个高频需求。 YearsS 直接拿 MonthsS(12) 乘倍数返回年秒数，默认 multiple=1 即单年。ReturnYear / ReturnMonth / ReturnDay 都靠 TimeToStruct 把 datetime 填进 MqlDateTime 结构，再取 year、mon、day 字段——注意 MT5 里 mon 是 1~12，不是 0 起始。 TimeMinusOffset 与 TimePlusOffset 是一对反向操作：入参是标准 datetime 和秒级偏移，前者减后者加，返回新的 datetime。实盘里用来算 N 根 K 线前的触发时刻或跨时区对齐都顺手。 外汇与贵金属杠杆高、跳空频繁，用秒偏移回测历史时必须确认对应时段市场是否开盘，否则可能取到无效报价。
class="type">int CTimeManagement::YearsS(class="type">int multiple=class="num">1) { class="kw">return (MonthsS(class="num">12)*multiple); } class="type">int CTimeManagement::ReturnYear(class="type">class="kw">datetime time) { TimeToStruct(time,timeFormat); class="kw">return timeFormat.year; } class="type">int CTimeManagement::ReturnMonth(class="type">class="kw">datetime time) { TimeToStruct(time,timeFormat); class="kw">return timeFormat.mon; } class="type">int CTimeManagement::ReturnDay(class="type">class="kw">datetime time) { TimeToStruct(time,timeFormat); class="kw">return timeFormat.day; } class="type">class="kw">datetime CTimeManagement::TimeMinusOffset(class="type">class="kw">datetime standardtime,class="type">int timeoffset) { standardtime-=timeoffset; class="kw">return standardtime; } class="type">class="kw">datetime CTimeManagement::TimePlusOffset(class="type">class="kw">datetime standardtime,class="type">int timeoffset) { standardtime+=timeoffset; class="kw">return standardtime; }
EA卸载后自动还原图表外观
写 EA 时顺手改了图表配色和周期分隔线，万一哪天把 EA 从图上拖掉，MT5 默认不会帮你把绿底黑字之外的布局变回去。CChartProperties 这个类的设计意图很直接：在改动前把图表配置存进数组，EA 被移除时由析构函数一键恢复整数型属性原值。 它单一继承自 CSymbolProperties，再往上追一层是 CSymbolInfo，所以取品种背景色这类活儿能直接复用父类方法。声明里的 ChartFormat 结构存了 14 个 ulong 字段，覆盖 CHART_MODE、CHART_COLOR_BACKGROUND、CHART_COLOR_ASK 等到 CHART_FOREGROUND，基本把视觉相关的图表属性捞全了。 构造函数里会先用 ChartConfig 数组备份当前图表的全部整型属性；SetBackground 从 MarketWatch 取该品种背景色再写到图表；ChartConfigure 给 Chart 变量（ChartFormat 类型）赋自定义值；ChartSet 照着结构往图表写；最后析构函数恢复备份。外汇和贵金属波动快、跳空多，图表高亮区分买卖价与止损线，可能让你在消息行情中更快定位水平，但杠杆品种高风险，视觉优化不替代风控。 开 MT5 把下面代码存成 mqh 引用，挂 EA 前后切一次图表，就能验证卸载后配色是否回弹。
class="macro">#include "SymbolProperties.mqh" class CChartProperties : class="kw">public CSymbolProperties { class="kw">private: class="kw">struct ChartFormat { class="type">ulong CHART_MODE;class=class="str">"cmt">//Chart Candle Mode class="type">ulong CHART_COLOR_BACKGROUND;class=class="str">"cmt">//Chart Background Color class="type">ulong CHART_COLOR_FOREGROUND;class=class="str">"cmt">//Chart Foreground Color class="type">ulong CHART_COLOR_CHART_LINE;class=class="str">"cmt">//Chart Line Color class="type">ulong CHART_COLOR_CANDLE_BEAR;class=class="str">"cmt">//Chart Bear Candle Color class="type">ulong CHART_COLOR_CHART_DOWN;class=class="str">"cmt">//Chart Down Candle Color class="type">ulong CHART_COLOR_CANDLE_BULL;class=class="str">"cmt">//Chart Bull Candle Color class="type">ulong CHART_COLOR_CHART_UP;class=class="str">"cmt">//Chart Up Candle Color class="type">ulong CHART_COLOR_ASK;class=class="str">"cmt">//Chart Ask Color class="type">ulong CHART_COLOR_BID;class=class="str">"cmt">//Chart Bid Color class="type">ulong CHART_COLOR_STOP_LEVEL;class=class="str">"cmt">//Chart Stoplevel Color class="type">ulong CHART_SHOW_PERIOD_SEP;class=class="str">"cmt">//Chart Show Period Separator class="type">ulong CHART_SCALE;class=class="str">"cmt">//Chart Scale class="type">ulong CHART_FOREGROUND;class=class="str">"cmt">//Chart Show Foreground }; };
「用类把图表属性一次性存回原样」
做价格行为分析时，脚本常要临时改图表显示（隐藏网格、关成交量、开一键交易），退出若不留痕，下一次手动调回来很烦。下面这段把图表整数类属性整体快照，构造时读、析构时写回。 核心是一个 65 长度的 ulong 数组 ChartConfig[65]，对应 ENUM_CHART_PROPERTY_INTEGER 的 0~64 共 65 个枚举值。构造函数里 for 循环调 ChartGetInteger(0, (ENUM_CHART_PROPERTY_INTEGER)i, 0, ChartConfig[i]) 把当前图表的 65 项整数属性全部存进数组；析构函数里反向用 ChartSetInteger 把存的值还原。 类里还留了 ChartConfigure() 和 ChartSet() 两个接口，分别负责设值和套用格式，外部只要调 ChartRefresh() 就会重跑 ChartConfigure()。这样你跑完剥头皮脚本，对象销毁后 MT5 图表自动回到跑之前的显示状态，不会留下半屏 OHLC 或偏移。 外汇和贵金属波动快、跳空频繁，图表显示项乱改可能让你漏看关键价位，用这类封装降低人为干扰概率。
class="type">ulong CHART_SHOW_ASK_LINE; class=class="str">"cmt">//Chart Show Ask Line class="type">ulong CHART_SHOW_BID_LINE; class=class="str">"cmt">//Chart Show Bid Line class="type">ulong CHART_SHOW_TRADE_LEVELS; class=class="str">"cmt">//Chart Show Trade Levels class="type">ulong CHART_SHOW_OHLC; class=class="str">"cmt">//Chart Show Open-High-Low-Close class="type">ulong CHART_SHOW_GRID; class=class="str">"cmt">//Chart Show Grid class="type">ulong CHART_SHOW_VOLUMES; class=class="str">"cmt">//Chart Show Volumes class="type">ulong CHART_AUTOSCROLL; class=class="str">"cmt">//Chart Auto Scroll class="type">class="kw">double CHART_SHIFT_SIZE; class=class="str">"cmt">//Chart Shift Size class="type">ulong CHART_SHIFT; class=class="str">"cmt">//Chart Shift class="type">ulong CHART_SHOW_ONE_CLICK; class=class="str">"cmt">//Chart One Click Trading }; class="type">ulong ChartConfig[class="num">65]; class=class="str">"cmt">//Array To Store Chart Properties class="type">void ChartSet(); class=class="str">"cmt">//Apply Chart format class="type">void ChartConfigure(); class=class="str">"cmt">//Set Chart Values ChartFormat Chart; class=class="str">"cmt">//Variable of type ChartFormat class="kw">public: CChartProperties(class="type">void); class=class="str">"cmt">//Constructor ~CChartProperties(class="type">void); class=class="str">"cmt">//Destructor class="type">void ChartRefresh() {ChartConfigure();} }; CChartProperties::CChartProperties(class="type">void) class=class="str">"cmt">//Class Constructor { for(class="type">int i=class="num">0;i<class="num">65;i++) class=class="str">"cmt">//Iterating through ENUM_CHART_PROPERTY_INTEGER Elements { ChartGetInteger(class="num">0,(ENUM_CHART_PROPERTY_INTEGER)i,class="num">0,ChartConfig[i]); class=class="str">"cmt">//Storing Chart values into ChartConfig array } ChartConfigure(); } CChartProperties::~CChartProperties(class="type">void) { for(class="type">int i=class="num">0;i<class="num">65;i++) class=class="str">"cmt">//Iterating through ENUM_CHART_PROPERTY_INTEGER Elements { ChartSetInteger(class="num">0,(ENUM_CHART_PROPERTY_INTEGER)i,class="num">0,ChartConfig[i]); class=class="str">"cmt">//Restoring Chart values from ChartConfig array } }
◍ 用类封装一键套用图表外观
在 MT5 里反复手动调图表颜色、网格、OHLC 显示，既耗时又容易在切换品种时丢失习惯视图。把图表属性收进一个 CChartProperties 类的 ChartSet() 方法，可以在 EA 初始化时一次性把当前 Chart 对象里的配置刷到 0 号图表（即当前主图）。 下面这段实现直接遍历了 23 个整数型属性和 1 个双精度属性，覆盖蜡烛模式、背景/前景色、多空蜡烛色、Bid/Ask 线、止损位色、周期分隔符、成交量等。注意 CHART_SHIFT_SIZE 必须用 ChartSetDouble 写，其余用 ChartSetInteger，混用会在编译期报错。 别把对象属性当默认值 Chart 对象里的字段若没在构造时显式赋值，MT5 会沿用终端全局设置；EA 换一台机器加载，图表可能突然变成白底黑字。建议把这套 ChartSet() 放在 OnInit 首行，保证每次加载视觉一致。 外汇与贵金属杠杆高、滑点大，自动化改图表属性虽不影响下单逻辑，但视图错乱可能让你误判支撑阻力，实盘前务必在策略测试器或模拟盘验证显示效果。
class="type">void CChartProperties::ChartSet() { ChartSetInteger(class="num">0,CHART_MODE,Chart.CHART_MODE);class=class="str">"cmt">//设置图表蜡烛模式 ChartSetInteger(class="num">0,CHART_COLOR_BACKGROUND,Chart.CHART_COLOR_BACKGROUND);class=class="str">"cmt">//设置图表背景色 ChartSetInteger(class="num">0,CHART_COLOR_FOREGROUND,Chart.CHART_COLOR_FOREGROUND);class=class="str">"cmt">//设置图表前景色 ChartSetInteger(class="num">0,CHART_COLOR_CHART_LINE,Chart.CHART_COLOR_CHART_LINE);class=class="str">"cmt">//设置图表轴线色 ChartSetInteger(class="num">0,CHART_COLOR_CANDLE_BEAR,Chart.CHART_COLOR_CANDLE_BEAR);class=class="str">"cmt">//设置阴线蜡烛色 ChartSetInteger(class="num">0,CHART_COLOR_CHART_DOWN,Chart.CHART_COLOR_CHART_DOWN);class=class="str">"cmt">//设置下跌蜡烛色 ChartSetInteger(class="num">0,CHART_COLOR_CANDLE_BULL,Chart.CHART_COLOR_CANDLE_BULL);class=class="str">"cmt">//设置阳线蜡烛色 ChartSetInteger(class="num">0,CHART_COLOR_CHART_UP,Chart.CHART_COLOR_CHART_UP);class=class="str">"cmt">//设置上涨蜡烛色 ChartSetInteger(class="num">0,CHART_COLOR_ASK,Chart.CHART_COLOR_ASK);class=class="str">"cmt">//设置Ask线色 ChartSetInteger(class="num">0,CHART_COLOR_BID,Chart.CHART_COLOR_BID);class=class="str">"cmt">//设置Bid线色 ChartSetInteger(class="num">0,CHART_COLOR_STOP_LEVEL,Chart.CHART_COLOR_STOP_LEVEL);class=class="str">"cmt">//设置止损位色 ChartSetInteger(class="num">0,CHART_FOREGROUND,Chart.CHART_FOREGROUND);class=class="str">"cmt">//设置图表是否置顶显示 ChartSetInteger(class="num">0,CHART_SHOW_ASK_LINE,Chart.CHART_SHOW_ASK_LINE);class=class="str">"cmt">//设置Ask线可见性 ChartSetInteger(class="num">0,CHART_SHOW_BID_LINE,Chart.CHART_SHOW_BID_LINE);class=class="str">"cmt">//设置Bid线可见性 ChartSetInteger(class="num">0,CHART_SHOW_PERIOD_SEP,Chart.CHART_SHOW_PERIOD_SEP);class=class="str">"cmt">//设置周期分隔符可见性 ChartSetInteger(class="num">0,CHART_SHOW_TRADE_LEVELS,Chart.CHART_SHOW_TRADE_LEVELS);class=class="str">"cmt">//设置交易水平线可见性 ChartSetInteger(class="num">0,CHART_SHOW_OHLC,Chart.CHART_SHOW_OHLC);class=class="str">"cmt">//设置OHLC可见性 ChartSetInteger(class="num">0,CHART_SHOW_GRID,Chart.CHART_SHOW_GRID);class=class="str">"cmt">//设置网格可见性 ChartSetInteger(class="num">0,CHART_SHOW_VOLUMES,Chart.CHART_SHOW_VOLUMES);class=class="str">"cmt">//设置成交量可见性 ChartSetInteger(class="num">0,CHART_SCALE,Chart.CHART_SCALE);class=class="str">"cmt">//设置图表缩放值 ChartSetInteger(class="num">0,CHART_AUTOSCROLL,Chart.CHART_AUTOSCROLL);class=class="str">"cmt">//设置自动滚动 ChartSetDouble(class="num">0,CHART_SHIFT_SIZE,Chart.CHART_SHIFT_SIZE);class=class="str">"cmt">//设置偏移尺寸 ChartSetInteger(class="num">0,CHART_SHIFT,Chart.CHART_SHIFT);class=class="str">"cmt">//设置图表偏移选项 }
用类封装把图表刷成黑白极简风
在 MT5 里做价格行为分析，很多交易者受不了默认图表的花花绿绿。把蜡烛阴阳线压成黑白色、关掉网格和成交量，视觉噪声会明显下降，扫单时更容易盯住实体与影线边界。 下面这段来自一个 CChartProperties 类的 ChartConfigure 方法，集中给当前图表（Chart 对象）赋值了一系列 CHART_* 属性。它先把模式锁成 CHART_CANDLES，背景取 SymbolBackground() 的返回值，前景、线、熊烛、Bid/Ask 线全设成 clrBlack，只有牛烛填 clrWhite——也就是经典的「黑底白阳、黑阴」反差配色。 顺手把 OHLC 报价条、网格、成交量全部关掉（都赋 false），保留周期分隔线和交易价位线（true）。CHART_SCALE 设为 3、SHIFT_SIZE 给 30 且开启右移，能让最新 K 线始终离开右边框约 30 像素，不会贴边。最后 CHART_SHOW_ONE_CLICK 设 false 关掉一键交易面板，避免误点，再调 ChartSet() 落盘。
…
ChartSetInteger(class="num">0,CHART_SHOW_ONE_CLICK,Chart.CHART_SHOW_ONE_CLICK);class=class="str">"cmt">//Set Chart One Click Trading } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Initialize Chart Properties | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">void CChartProperties::ChartConfigure(class="type">void) { Chart.CHART_MODE=(class="type">ulong)CHART_CANDLES;class=class="str">"cmt">//Assigning Chart Mode of CHART_CANDLES Chart.CHART_COLOR_BACKGROUND=class="type">ulong(SymbolBackground());class=class="str">"cmt">//Assigning Chart Background Color of Symbol&class="macro">#x27;s Background class="type">class="kw">color Chart.CHART_COLOR_FOREGROUND=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Foreground Color of clrBalck(Black class="type">class="kw">color) Chart.CHART_COLOR_CHART_LINE=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Line Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_CANDLE_BEAR=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Bear Candle Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_CHART_DOWN=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Down Candle Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_CANDLE_BULL=(class="type">ulong)clrWhite;class=class="str">"cmt">//Assigning Chart Bull Candle Color of clrWhite(White class="type">class="kw">color) Chart.CHART_COLOR_CHART_UP=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Up Candle Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_ASK=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Ask Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_BID=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Bid Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_STOP_LEVEL=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Stop Level Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_FOREGROUND=(class="type">ulong)false;class=class="str">"cmt">//Assigning Chart Foreground Boolean Value of &class="macro">#x27;false&class="macro">#x27; Chart.CHART_SHOW_ASK_LINE=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Ask Line Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHOW_BID_LINE=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Bid Line Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHOW_PERIOD_SEP=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Period Separator Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHOW_TRADE_LEVELS=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Trade Levels Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHOW_OHLC=(class="type">ulong)false;class=class="str">"cmt">//Assigning Chart Open-High-Low-Close Boolean Value of &class="macro">#x27;false&class="macro">#x27; Chart.CHART_SHOW_GRID=(class="type">ulong)false;class=class="str">"cmt">//Assigning Chart Grid Boolean Value of &class="macro">#x27;false&class="macro">#x27; Chart.CHART_SHOW_VOLUMES=(class="type">ulong)false;class=class="str">"cmt">//Assigning Chart Volumes Boolean Value of &class="macro">#x27;false&class="macro">#x27; Chart.CHART_SCALE=(class="type">ulong)class="num">3;class=class="str">"cmt">//Assigning Chart Scale Boolean Value of &class="macro">#x27;class="num">3&class="macro">#x27; Chart.CHART_AUTOSCROLL=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Auto Scroll Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHIFT_SIZE=class="num">30;class=class="str">"cmt">//Assigning Chart Shift Size Value of &class="macro">#x27;class="num">30&class="macro">#x27; Chart.CHART_SHIFT=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Shift Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHOW_ONE_CLICK=class="type">ulong(false);class=class="str">"cmt">//Assigning Chart One Click Trading a value of &class="macro">#x27;false&class="macro">#x27; ChartSet();class=class="str">"cmt">//Calling Function to set chart format
「把图表外观塞进一个结构体」
做批量图表管理时，最怕每次改属性都写一长串 ChartSetInteger/ChartGetInteger。把常用图表属性收进一个结构体，后面存模板、还原模板、A/B 对比都只需操作一块内存。 下面这段结构体定义了 24 个字段，覆盖蜡烛模式、背景/前景色、涨跌蜡烛色、Ask/Bid 线、止损位线、周期分隔符、缩放、网格、成交量等。注意 CHART_SHIFT_SIZE 用的是 double，其余大多是 ulong，因为 MT5 图表属性枚举本质是整数 ID。 实盘外接小布盯盘时，你可以先读出现图表的这 24 个值存盘，再切到自定义皮肤；看完价格行为想一键还原，直接把结构体写回即可，不用手动点 20 多次鼠标。外汇与贵金属波动剧烈，改图表视觉参数不影响风险，但别误以为换皮肤能改信号质量。
class="kw">struct ChartFormat { class="type">ulong CHART_MODE;class=class="str">"cmt">//Chart Candle Mode class="type">ulong CHART_COLOR_BACKGROUND;class=class="str">"cmt">//Chart Background Color class="type">ulong CHART_COLOR_FOREGROUND;class=class="str">"cmt">//Chart Foreground Color class="type">ulong CHART_COLOR_CHART_LINE;class=class="str">"cmt">//Chart Line Color class="type">ulong CHART_COLOR_CANDLE_BEAR;class=class="str">"cmt">//Chart Bear Candle Color class="type">ulong CHART_COLOR_CHART_DOWN;class=class="str">"cmt">//Chart Down Candle Color class="type">ulong CHART_COLOR_CANDLE_BULL;class=class="str">"cmt">//Chart Bull Candle Color class="type">ulong CHART_COLOR_CHART_UP;class=class="str">"cmt">//Chart Up Candle Color class="type">ulong CHART_COLOR_ASK;class=class="str">"cmt">//Chart Ask Color class="type">ulong CHART_COLOR_BID;class=class="str">"cmt">//Chart Bid Color class="type">ulong CHART_COLOR_STOP_LEVEL;class=class="str">"cmt">//Chart Stoplevel Color class="type">ulong CHART_SHOW_PERIOD_SEP;class=class="str">"cmt">//Chart Show Period Separator class="type">ulong CHART_SCALE;class=class="str">"cmt">//Chart Scale class="type">ulong CHART_FOREGROUND;class=class="str">"cmt">//Chart Show Foreground class="type">ulong CHART_SHOW_ASK_LINE;class=class="str">"cmt">//Chart Show Ask Line class="type">ulong CHART_SHOW_BID_LINE;class=class="str">"cmt">//Chart Show Bid Line class="type">ulong CHART_SHOW_TRADE_LEVELS;class=class="str">"cmt">//Chart Show Trade Levels class="type">ulong CHART_SHOW_OHLC;class=class="str">"cmt">//Chart Show Open-High-Low-Close class="type">ulong CHART_SHOW_GRID;class=class="str">"cmt">//Chart Show Grid class="type">ulong CHART_SHOW_VOLUMES;class=class="str">"cmt">//Chart Show Volumes class="type">ulong CHART_AUTOSCROLL;class=class="str">"cmt">//Chart Auto Scroll class="type">class="kw">double CHART_SHIFT_SIZE;class=class="str">"cmt">//Chart Shift Size class="type">ulong CHART_SHIFT;class=class="str">"cmt">//Chart Shift };
◍ 用类构造函数批量接管图表属性
MQL5 里把图表视觉参数集中管理，常见做法是写一个 CChartProperties 类，在构造函数里一次性把当前图表整数型属性读进数组。下面这段代码用长度为 65 的 ulong 数组 ChartConfig 承接 ENUM_CHART_PROPERTY_INTEGER 的全部元素，遍历下标 0 到 64 调用 ChartGetInteger 完成快照。 构造函数末尾顺手调了 ChartConfigure()，后者直接给 Chart 对象逐条赋值：K线模式锁死 CHART_CANDLES，背景色取 SymbolBackground() 返回值，前景、轴线、熊线、卖价买价线全压成 clrBlack，只有 bull 蜡烛填 clrWhite。 这种写法在 MT5 上跑起来，意味着你每次 new 这个类，图表会被强制刷成黑白极简风，且 Ask/Bid 线、周期分隔线、交易水平线均强制显示。外汇与贵金属波动剧烈、杠杆风险高，视觉统一有助于减少误读，但参数覆盖后原图表风格会丢失，实盘前建议在模拟账户验证。
class="type">ulong CHART_SHOW_ONE_CLICK;class=class="str">"cmt">//Chart One Click Trading }; class="type">ulong ChartConfig[class="num">65];class=class="str">"cmt">//Array To Store Chart Properties CChartProperties::CChartProperties(class="type">void)class=class="str">"cmt">//Class Constructor { for(class="type">int i=class="num">0;i<class="num">65;i++)class=class="str">"cmt">//Iterating through ENUM_CHART_PROPERTY_INTEGER Elements { ChartGetInteger(class="num">0,(ENUM_CHART_PROPERTY_INTEGER)i,class="num">0,ChartConfig[i]);class=class="str">"cmt">//Storing Chart values into ChartConfig array } ChartConfigure(); } class="type">void CChartProperties::ChartConfigure(class="type">void) { Chart.CHART_MODE=(class="type">ulong)CHART_CANDLES;class=class="str">"cmt">//Assigning Chart Mode of CHART_CANDLES Chart.CHART_COLOR_BACKGROUND=class="type">ulong(SymbolBackground());class=class="str">"cmt">//Assigning Chart Background Color of Symbol&class="macro">#x27;s Background class="type">class="kw">color Chart.CHART_COLOR_FOREGROUND=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Foreground Color of clrBalck(Black class="type">class="kw">color) Chart.CHART_COLOR_CHART_LINE=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Line Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_CANDLE_BEAR=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Bear Candle Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_CHART_DOWN=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Down Candle Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_CANDLE_BULL=(class="type">ulong)clrWhite;class=class="str">"cmt">//Assigning Chart Bull Candle Color of clrWhite(White class="type">class="kw">color) Chart.CHART_COLOR_CHART_UP=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Up Candle Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_ASK=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Ask Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_BID=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Bid Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_STOP_LEVEL=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Stop Level Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_FOREGROUND=(class="type">ulong)false;class=class="str">"cmt">//Assigning Chart Foreground Boolean Value of &class="macro">#x27;false&class="macro">#x27; Chart.CHART_SHOW_ASK_LINE=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Ask Line Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHOW_BID_LINE=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Bid Line Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHOW_PERIOD_SEP=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Period Separator Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHOW_TRADE_LEVELS=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Trade Levels Boolean Value of &class="macro">#x27;true&class="macro">#x27;
ulong CHART_SHOW_ONE_CLICK; 声明一键交易属性变量，注释标明用途。
ulong ChartConfig[65]; 开 65 格数组，对应全部整数型图表属性枚举。
CChartProperties::CChartProperties(void) 类构造函数入口。
for(int i=0;i<65;i++) 从 0 循环到 64。
ChartGetInteger(0,(ENUM_CHART_PROPERTY_INTEGER)i,0,ChartConfig[i]); 把当前图表（句柄0）第 i 个整数属性读入数组。
ChartConfigure(); 构造函数收尾调用配置函数。
Chart.CHART_MODE=(ulong)CHART_CANDLES; 图表切蜡烛图。
Chart.CHART_COLOR_BACKGROUND=ulong(SymbolBackground()); 背景色跟随品种设置。
Chart.CHART_COLOR_FOREGROUND=(ulong)clrBlack; 前景黑。
Chart.CHART_COLOR_CHART_LINE=(ulong)clrBlack; 图表线黑。
Chart.CHART_COLOR_CANDLE_BEAR=(ulong)clrBlack; 熊烛黑。
Chart.CHART_COLOR_CHART_DOWN=(ulong)clrBlack; 下跌元素黑。
Chart.CHART_COLOR_CANDLE_BULL=(ulong)clrWhite; 牛烛白。
Chart.CHART_COLOR_CHART_UP=(ulong)clrBlack; 上涨元素黑。
Chart.CHART_COLOR_ASK=(ulong)clrBlack; 卖价线黑。
Chart.CHART_COLOR_BID=(ulong)clrBlack; 买价线黑。
Chart.CHART_COLOR_STOP_LEVEL=(ulong)clrBlack; 止损位黑。
Chart.CHART_FOREGROUND=(ulong)false; 禁用前景置顶。
Chart.CHART_SHOW_ASK_LINE=(ulong)true; 显示卖价线。
Chart.CHART_SHOW_BID_LINE=(ulong)true; 显示买价线。
Chart.CHART_SHOW_PERIOD_SEP=(ulong)true; 显示周期分割。
Chart.CHART_SHOW_TRADE_LEVELS=(ulong)true; 显示交易水平。
class="type">ulong CHART_SHOW_ONE_CLICK;class=class="str">"cmt">//Chart One Click Trading }; class="type">ulong ChartConfig[class="num">65];class=class="str">"cmt">//Array To Store Chart Properties CChartProperties::CChartProperties(class="type">void)class=class="str">"cmt">//Class Constructor { for(class="type">int i=class="num">0;i<class="num">65;i++)class=class="str">"cmt">//Iterating through ENUM_CHART_PROPERTY_INTEGER Elements { ChartGetInteger(class="num">0,(ENUM_CHART_PROPERTY_INTEGER)i,class="num">0,ChartConfig[i]);class=class="str">"cmt">//Storing Chart values into ChartConfig array } ChartConfigure(); } class="type">void CChartProperties::ChartConfigure(class="type">void) { Chart.CHART_MODE=(class="type">ulong)CHART_CANDLES;class=class="str">"cmt">//Assigning Chart Mode of CHART_CANDLES Chart.CHART_COLOR_BACKGROUND=class="type">ulong(SymbolBackground());class=class="str">"cmt">//Assigning Chart Background Color of Symbol&class="macro">#x27;s Background class="type">class="kw">color Chart.CHART_COLOR_FOREGROUND=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Foreground Color of clrBalck(Black class="type">class="kw">color) Chart.CHART_COLOR_CHART_LINE=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Line Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_CANDLE_BEAR=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Bear Candle Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_CHART_DOWN=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Down Candle Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_CANDLE_BULL=(class="type">ulong)clrWhite;class=class="str">"cmt">//Assigning Chart Bull Candle Color of clrWhite(White class="type">class="kw">color) Chart.CHART_COLOR_CHART_UP=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Up Candle Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_ASK=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Ask Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_BID=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Bid Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_COLOR_STOP_LEVEL=(class="type">ulong)clrBlack;class=class="str">"cmt">//Assigning Chart Stop Level Color of clrBlack(Black class="type">class="kw">color) Chart.CHART_FOREGROUND=(class="type">ulong)false;class=class="str">"cmt">//Assigning Chart Foreground Boolean Value of &class="macro">#x27;false&class="macro">#x27; Chart.CHART_SHOW_ASK_LINE=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Ask Line Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHOW_BID_LINE=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Bid Line Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHOW_PERIOD_SEP=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Period Separator Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHOW_TRADE_LEVELS=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Trade Levels Boolean Value of &class="macro">#x27;true&class="macro">#x27;
用类封装把图表属性一次性写进MT5
在自定义类里集中管理图表外观，比每次手动点界面更高效，也方便多图表同步。下面这段把 OHLC 标签、网格、成交量全关掉，缩放级别设成 3，并开启自动滚动与右侧留白偏移 30 像素，顺手禁用了一键交易面板。 CHART_SHIFT_SIZE=30 配合 CHART_SHIFT=true，能让最新 K 线固定在离右边界 30 像素处，盯盘时价格动作不会被挤到边缘。关掉 CHART_SHOW_OHLC 和 CHART_SHOW_VOLUMES 后，裸 K 视野更干净，适合做价格行为读取。 ChartSet() 函数里用 ChartSetInteger(0,...) 逐条把成员变量刷到当前图表（图表 ID 用 0 即当前）。颜色类属性如 CHART_COLOR_CANDLE_BULL 也在这里统一下发，改配色只需动类成员不用碰调用处。 复制进 EA 或脚本的 CChartProperties 类，编译后调用 ChartSet() 即可在任意品种图表验证这套极简布局；外汇与贵金属波动剧烈，高风险下清爽视图有助于减少误读。
Chart.CHART_SHOW_OHLC=(class="type">ulong)false;class=class="str">"cmt">//Assigning Chart Open-High-Low-Close Boolean Value of &class="macro">#x27;false&class="macro">#x27; Chart.CHART_SHOW_GRID=(class="type">ulong)false;class=class="str">"cmt">//Assigning Chart Grid Boolean Value of &class="macro">#x27;false&class="macro">#x27; Chart.CHART_SHOW_VOLUMES=(class="type">ulong)false;class=class="str">"cmt">//Assigning Chart Volumes Boolean Value of &class="macro">#x27;false&class="macro">#x27; Chart.CHART_SCALE=(class="type">ulong)class="num">3;class=class="str">"cmt">//Assigning Chart Scale Boolean Value of &class="macro">#x27;class="num">3&class="macro">#x27; Chart.CHART_AUTOSCROLL=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Auto Scroll Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHIFT_SIZE=class="num">30;class=class="str">"cmt">//Assigning Chart Shift Size Value of &class="macro">#x27;class="num">30&class="macro">#x27; Chart.CHART_SHIFT=(class="type">ulong)true;class=class="str">"cmt">//Assigning Chart Shift Boolean Value of &class="macro">#x27;true&class="macro">#x27; Chart.CHART_SHOW_ONE_CLICK=class="type">ulong(false);class=class="str">"cmt">//Assigning Chart One Click Trading a value of &class="macro">#x27;false&class="macro">#x27; ChartSet();class=class="str">"cmt">//Calling Function to set chart format } class="type">void CChartProperties::ChartSet() { ChartSetInteger(class="num">0,CHART_MODE,Chart.CHART_MODE);class=class="str">"cmt">//Set Chart Candle Mode ChartSetInteger(class="num">0,CHART_COLOR_BACKGROUND,Chart.CHART_COLOR_BACKGROUND);class=class="str">"cmt">//Set Chart Background Color ChartSetInteger(class="num">0,CHART_COLOR_FOREGROUND,Chart.CHART_COLOR_FOREGROUND);class=class="str">"cmt">//Set Chart Foreground Color ChartSetInteger(class="num">0,CHART_COLOR_CHART_LINE,Chart.CHART_COLOR_CHART_LINE);class=class="str">"cmt">//Set Chart Line Color ChartSetInteger(class="num">0,CHART_COLOR_CANDLE_BEAR,Chart.CHART_COLOR_CANDLE_BEAR);class=class="str">"cmt">//Set Chart Bear Candle Color ChartSetInteger(class="num">0,CHART_COLOR_CHART_DOWN,Chart.CHART_COLOR_CHART_DOWN);class=class="str">"cmt">//Set Chart Down Candle Color ChartSetInteger(class="num">0,CHART_COLOR_CANDLE_BULL,Chart.CHART_COLOR_CANDLE_BULL);class=class="str">"cmt">//Set Chart Bull Candle Color ChartSetInteger(class="num">0,CHART_COLOR_CHART_UP,Chart.CHART_COLOR_CHART_UP);class=class="str">"cmt">//Set Chart Up Candle Color ChartSetInteger(class="num">0,CHART_COLOR_ASK,Chart.CHART_COLOR_ASK);class=class="str">"cmt">//Set Chart Ask Color ChartSetInteger(class="num">0,CHART_COLOR_BID,Chart.CHART_COLOR_BID);class=class="str">"cmt">//Set Chart Bid Color ChartSetInteger(class="num">0,CHART_COLOR_STOP_LEVEL,Chart.CHART_COLOR_STOP_LEVEL);class=class="str">"cmt">//Set Chart Stop Level Color
「用代码接管图表外观与析构还原」
在 MT5 里想批量控制当前图表的可见元素，直接调 ChartSetInteger 往 0 号图表句柄写属性最快。下面这段把前景、买卖价线、周期分隔、交易水平、OHLC、网格、成交量、缩放、自动滚动、右移、一键交易等一次性设完，省得手动点界面。 ChartSetInteger(0,CHART_FOREGROUND,Chart.CHART_FOREGROUND); // 设图表是否置顶前景 ChartSetInteger(0,CHART_SHOW_ASK_LINE,Chart.CHART_SHOW_ASK_LINE); // 控制卖价线显示 ChartSetInteger(0,CHART_SHOW_BID_LINE,Chart.CHART_SHOW_BID_LINE); // 控制买价线显示 ChartSetInteger(0,CHART_SHOW_PERIOD_SEP,Chart.CHART_SHOW_PERIOD_SEP);// 周期分隔符开关 ChartSetInteger(0,CHART_SHOW_TRADE_LEVELS,Chart.CHART_SHOW_TRADE_LEVELS);// 持仓挂单水平线 ChartSetInteger(0,CHART_SHOW_OHLC,Chart.CHART_SHOW_OHLC); // 左上角OHLC报价块 ChartSetInteger(0,CHART_SHOW_GRID,Chart.CHART_SHOW_GRID); // 网格线 ChartSetInteger(0,CHART_SHOW_VOLUMES,Chart.CHART_SHOW_VOLUMES); // 成交量柱 ChartSetInteger(0,CHART_SCALE,Chart.CHART_SCALE); // 缩放档位 ChartSetInteger(0,CHART_AUTOSCROLL,Chart.CHART_AUTOSCROLL); // 自动滚动跟随 ChartSetDouble(0,CHART_SHIFT_SIZE,Chart.CHART_SHIFT_SIZE); // 右移留白像素 ChartSetInteger(0,CHART_SHIFT,Chart.CHART_SHIFT); // 右移开关 ChartSetInteger(0,CHART_SHOW_ONE_CLICK,Chart.CHART_SHOW_ONE_CLICK); // 一键交易面板 析构函数里用 for(int i=0;i<65;i++) 遍历 ENUM_CHART_PROPERTY_INTEGER 全部 65 个整型属性，从 ChartConfig 数组读回原值写回图表。也就是说这个类在销毁时会把图表还原到你接管前的状态，做 EA 面板时不怕弄脏用户图表。 外汇和贵金属波动大、杠杆高，这类图表脚本只改显示不影响下单逻辑，但实盘前先在策略测试器或模拟盘验证还原是否完整，避免关 EA 后界面残留。
ChartSetInteger(class="num">0,CHART_FOREGROUND,Chart.CHART_FOREGROUND);class=class="str">"cmt">//Set if Chart is in Foreground Visibility ChartSetInteger(class="num">0,CHART_SHOW_ASK_LINE,Chart.CHART_SHOW_ASK_LINE);class=class="str">"cmt">//Set Chart Ask Line Visibility ChartSetInteger(class="num">0,CHART_SHOW_BID_LINE,Chart.CHART_SHOW_BID_LINE);class=class="str">"cmt">//Set Chart Bid Line Visibility ChartSetInteger(class="num">0,CHART_SHOW_PERIOD_SEP,Chart.CHART_SHOW_PERIOD_SEP);class=class="str">"cmt">//Set Chart Period Separator Visibility ChartSetInteger(class="num">0,CHART_SHOW_TRADE_LEVELS,Chart.CHART_SHOW_TRADE_LEVELS);class=class="str">"cmt">//Set Chart Trade Levels Visibility ChartSetInteger(class="num">0,CHART_SHOW_OHLC,Chart.CHART_SHOW_OHLC);class=class="str">"cmt">//Set Chart Open-High-Low-Close Visibility ChartSetInteger(class="num">0,CHART_SHOW_GRID,Chart.CHART_SHOW_GRID);class=class="str">"cmt">//Set Chart Grid Visibility ChartSetInteger(class="num">0,CHART_SHOW_VOLUMES,Chart.CHART_SHOW_VOLUMES);class=class="str">"cmt">//Set Chart Volumes Visibility ChartSetInteger(class="num">0,CHART_SCALE,Chart.CHART_SCALE);class=class="str">"cmt">//Set Chart Scale Value ChartSetInteger(class="num">0,CHART_AUTOSCROLL,Chart.CHART_AUTOSCROLL);class=class="str">"cmt">//Set Chart Auto Scroll Option ChartSetDouble(class="num">0,CHART_SHIFT_SIZE,Chart.CHART_SHIFT_SIZE);class=class="str">"cmt">//Set Chart Shift Size Value ChartSetInteger(class="num">0,CHART_SHIFT,Chart.CHART_SHIFT);class=class="str">"cmt">//Set Chart Shift Option ChartSetInteger(class="num">0,CHART_SHOW_ONE_CLICK,Chart.CHART_SHOW_ONE_CLICK);class=class="str">"cmt">//Set Chart One Click Trading } CChartProperties::~CChartProperties(class="type">void) { for(class="type">int i=class="num">0;i<class="num">65;i++)class=class="str">"cmt">//Iterating through ENUM_CHART_PROPERTY_INTEGER Elements { ChartSetInteger(class="num">0,(ENUM_CHART_PROPERTY_INTEGER)i,class="num">0,ChartConfig[i]);class=class="str">"cmt">//Restoring Chart values from ChartConfig array } }
◍ K线属性类的继承与取价接口
在 MT5 的自定义分析框架里，CCandleProperties 不是凭空写的类，它走的是多层级继承：先吃下 CChartProperties，再经由 CSymbolProperties、CSymbolInfo 把品种维度接进来，同时内部持有 CTimeManagement 实例管时间。这样设计的好处是，调一根 K 线的开高低收不需要每次重连品种上下文，类自己就带着符号与周期状态。 对外暴露的取价方法都给了默认值：CandleIndex 定位第几根，Period 默认 PERIOD_CURRENT，SYMBOL 默认 NULL 表示用当前图品种。Open() 里先跑 SetSymbolName(SYMBOL) 确立品种，再调 iOpen(GetSymbolName(),Period,CandleIndex)，失败返回 0——这种写法在跨品种扫描时容易把「取不到」和「价格为 0」混为一谈，实盘前最好加日志区分。 还有个 IsLargerThanPreviousAndNext(datetime CandleTime,int Offset,string SYMBOL)，用来判断某根 K 线实体是否同时大于相邻两根，做吞没或异常放量形态筛选用得上。外汇与贵金属波动受杠杆与消息影响大，这类形态只是概率信号，真上线前请在策略测试器里用不同品种跑一遍。
class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| NewsTrading | class=class="str">"cmt">//| Copyright class="num">2024, MetaQuotes Ltd. | class=class="str">"cmt">//| [MQL5官方文档] | class=class="str">"cmt">//+------------------------------------------------------------------+ class="macro">#include "TimeManagement.mqh" class="macro">#include "ChartProperties.mqh" class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|CandleProperties class | class=class="str">"cmt">//+------------------------------------------------------------------+ class CCandleProperties : class="kw">public CChartProperties { class="kw">private: CTimeManagement Time; class="kw">public: class="type">class="kw">double Open(class="type">int CandleIndex,ENUM_TIMEFRAMES Period=PERIOD_CURRENT,class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Candle Open-Price class="type">class="kw">double Close(class="type">int CandleIndex,ENUM_TIMEFRAMES Period=PERIOD_CURRENT,class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Candle Close-Price class="type">class="kw">double High(class="type">int CandleIndex,ENUM_TIMEFRAMES Period=PERIOD_CURRENT,class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Candle High-Price class="type">class="kw">double Low(class="type">int CandleIndex,ENUM_TIMEFRAMES Period=PERIOD_CURRENT,class="type">class="kw">string SYMBOL=NULL);class=class="str">"cmt">//Retrieve Candle Low-Price class="type">bool IsLargerThanPreviousAndNext(class="type">class="kw">datetime CandleTime,class="type">int Offset,class="type">class="kw">string SYMBOL);class=class="str">"cmt">//Determine if one candle is larger than two others }; class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Candle Open-Price | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">double CCandleProperties::Open(class="type">int CandleIndex,ENUM_TIMEFRAMES Period=PERIOD_CURRENT,class="type">class="kw">string SYMBOL=NULL) { class="kw">return (SetSymbolName(SYMBOL))?iOpen(GetSymbolName(),Period,CandleIndex):class="num">0;class=class="str">"cmt">//class="kw">return candle open price } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Candle Close-Price | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">double CCandleProperties::Close(class="type">int CandleIndex,ENUM_TIMEFRAMES Period=PERIOD_CURRENT,class="type">class="kw">string SYMBOL=NULL) {
用K线高度比对锁定异动发生柱
在封装好的蜡烛属性类里，Close、High、Low 三个方法都走同一套路：先 SetSymbolName 切换品种上下文，成功才调用 iClose / iHigh / iLow 取数，失败直接返 0。这样写能让后续逻辑不用反复传 SYMBOL 参数，调用时只给 CandleIndex 和 Period 就行。 真正有实战价值的是 IsLargerThanPreviousAndNext 这个方法。它固定在 M15 周期上，用 iBarShift 把传入的 CandleTime 以及加减 Offset 后的两个时间点转成柱索引，再分别算三根 M15 蜡烛的实体加影线高度（High 减 Low）。 判定逻辑很直接：若中间那根 CandleHeight 同时大于前后两根，就返回 true，倾向认为该柱对应了新闻事件冲击；否则返回 false。你在 MT5 里把 Offset 设成 1 就是比相邻柱，设成 4 就能框出前后各 1 小时窗口（M15 下 4 根 = 1 小时），调这个参数就能改变“异动”的灵敏度。 外汇与贵金属受新闻跳空影响明显，这类识别只是概率性标记，实盘需结合风控，杠杆品种高风险。
class="kw">return (SetSymbolName(SYMBOL))?iClose(GetSymbolName(),Period,CandleIndex):class="num">0;class=class="str">"cmt">//class="kw">return candle close price } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Candle High-Price | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">double CCandleProperties::High(class="type">int CandleIndex,ENUM_TIMEFRAMES Period=PERIOD_CURRENT,class="type">class="kw">string SYMBOL=NULL) { class="kw">return (SetSymbolName(SYMBOL))?iHigh(GetSymbolName(),Period,CandleIndex):class="num">0;class=class="str">"cmt">//class="kw">return candle high price } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Retrieve Candle Low-Price | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">double CCandleProperties::Low(class="type">int CandleIndex,ENUM_TIMEFRAMES Period=PERIOD_CURRENT,class="type">class="kw">string SYMBOL=NULL) { class="kw">return (SetSymbolName(SYMBOL))?iLow(GetSymbolName(),Period,CandleIndex):class="num">0;class=class="str">"cmt">//class="kw">return candle low price } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Determine if one candle is larger than two others | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">bool CCandleProperties::IsLargerThanPreviousAndNext(class="type">class="kw">datetime CandleTime,class="type">int Offset,class="type">class="kw">string SYMBOL) { class="type">int CandleIndex = iBarShift(SYMBOL,PERIOD_M15,CandleTime);class=class="str">"cmt">//Assign candle index of candletime class=class="str">"cmt">//--Assign candle index of candletime minus time offset class="type">int CandleIndexMinusOffset = iBarShift(SYMBOL,PERIOD_M15,Time.TimeMinusOffset(CandleTime,Offset)); class=class="str">"cmt">//--Assign candle index of candletime plus time offset class="type">int CandleIndexPlusOffset = iBarShift(SYMBOL,PERIOD_M15,Time.TimePlusOffset(CandleTime,Offset)); class=class="str">"cmt">//--Assign height of M15 candletime in pips class="type">class="kw">double CandleHeight = High(CandleIndex,PERIOD_M15,SYMBOL)-Low(CandleIndex,PERIOD_M15,SYMBOL); class=class="str">"cmt">//--Assign height of M15 candletime minus offset in Pips class="type">class="kw">double CandleHeightMinusOffset = High(CandleIndexMinusOffset,PERIOD_M15,SYMBOL)-Low(CandleIndexMinusOffset,PERIOD_M15,SYMBOL); class=class="str">"cmt">//--Assign height of M15 candletime plus offset in Pips class="type">class="kw">double CandleHeightPlusOffset = High(CandleIndexPlusOffset,PERIOD_M15,SYMBOL)-Low(CandleIndexPlusOffset,PERIOD_M15,SYMBOL); class=class="str">"cmt">//--Determine if candletime height is greater than candletime height minus offset and candletime height plus offset if(CandleHeight>CandleHeightMinusOffset&&CandleHeight>CandleHeightPlusOffset) { class="kw">return true;class=class="str">"cmt">//Candletime is likely when the news event occured } class="kw">return false;class=class="str">"cmt">//Candletime is unlikely when the real news data was released } class=class="str">"cmt">//+------------------------------------------------------------------+
「图表对象的创建与回收机制」
在 MT5 的 EA 或指标工程里，把图表对象的管理单独抽成一个类会更干净。CObjectProperties 通过多层级继承拿到了 CChartProperties、CSymbolProperties 乃至 CSymbolInfo 的能力，本身只专注一件事：生成对象、登记对象、删除对象。 它内部用 ObjStruct 结构（含 ChartId 与 Name 两个字段）的 Objects[] 动态数组，记录每一次创建动作。AddObj 私有方法在每次画完图后把图表 ID 和对象名压进数组，ArrayResize 的步长设为当前大小加 2，意味着对象数增长时内存按 2 个为单位预留，避免频繁重分配。 对外暴露的三个公开方法覆盖了常见标注需求：Square 画可自定义属性的矩形（用于框选波段或风险区），TextObj 在指定角落贴文字（默认左上、字号 10），EventObj 打经济事件标记（带描述与发生时间）。析构函数 ~CObjectProperties 会遍历 Objects 数组把所有建过的对象清掉，所以一个实例生命周期结束，图表不会留垃圾。 外汇与贵金属杠杆高、滑点随机，任何对象标注都只是辅助参考，不代表方向判断。把下面这段核心登记逻辑直接拷进你的 .mqh 验证，看 Objects.Size() 是否随画图次数线性增加。
class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| NewsTrading | class=class="str">"cmt">//| Copyright class="num">2024, MetaQuotes Ltd. | class=class="str">"cmt">//| [MQL5官方文档] | class=class="str">"cmt">//+------------------------------------------------------------------+ class="macro">#include "ChartProperties.mqh" class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|ObjectProperties class | class=class="str">"cmt">//+------------------------------------------------------------------+ class CObjectProperties:class="kw">public CChartProperties { class="kw">private: class=class="str">"cmt">//Simple chart objects structure class="kw">struct ObjStruct { class="type">long ChartId; class="type">class="kw">string Name; } Objects[];class=class="str">"cmt">//ObjStruct variable array class=class="str">"cmt">//-- Add chart object to Objects array class="type">void AddObj(class="type">long chart_id,class="type">class="kw">string name) { ArrayResize(Objects,Objects.Size()+class="num">1,Objects.Size()+class="num">2); Objects[Objects.Size()-class="num">1].ChartId=chart_id; Objects[Objects.Size()-class="num">1].Name=name; } class="kw">public: CObjectProperties(class="type">void) {}class=class="str">"cmt">//Class constructor class=class="str">"cmt">//-- Create Rectangle chart object class="type">void Square(class="type">long chart_ID,class="type">class="kw">string name,class="type">int x_coord,class="type">int y_coord,class="type">int width,class="type">int height,ENUM_ANCHOR_POINT Anchor); class=class="str">"cmt">//-- Create text chart object class="type">void TextObj(class="type">long chartID,class="type">class="kw">string name,class="type">class="kw">string text,class="type">int x_coord,class="type">int y_coord, ENUM_BASE_CORNER Corner=CORNER_LEFT_UPPER,class="type">int fontsize=class="num">10); class=class="str">"cmt">//-- Create Event object class="type">void EventObj(class="type">long chartID,class="type">class="kw">string name,class="type">class="kw">string description,class="type">class="kw">datetime eventdate); class=class="str">"cmt">//-- Class destructor removes all chart objects created previously ~CObjectProperties(class="type">void) { for(class="type">uint i=class="num">0;i<Objects.Size();i++)
◍ 用矩形标签在图表上画可控色块
在 MT5 自定义指标或 EA 里，想在主图固定坐标叠一个纯视觉色块（比如高亮某根 K 线区域），用 OBJ_RECTANGLE_LABEL 比画趋势线更直接，它按像素距离定位、不随价格轴缩放。 下面这段 CObjectProperties::Square 方法先把同名旧对象删掉，避免重复堆叠；随后用 ObjectCreate 在 sub_window=0（主图）创建标签矩形，并把图表 ID 与名称登记进内部数组方便统一回收。 坐标与尺寸全部走 OBJPROP_XDISTANCE / YDISTANCE / XSIZE / YSIZE 四个整数属性，x、y 以图表左上角为原点，width、height 是像素值——这意味着你改一个参数就能让色块精确罩住想标记的区域，外汇与贵金属行情跳空时这种固定像素框不会变形，但这类品种杠杆高、滑点大，实盘前务必在策略测试器里验证锚点逻辑。 背景色写死 clrBlack、边框 clrRed 且线宽 0，hidden=true 让对象不出现在终端对象列表里，适合做‘安静的’界面提示层。
class="type">void CObjectProperties::Square(class="type">long chart_ID,class="type">class="kw">string name,class="type">int x_coord,class="type">int y_coord,class="type">int width,class="type">int height,ENUM_ANCHOR_POINT Anchor) { const class="type">int sub_window=class="num">0; class=class="str">"cmt">// subwindow index const class="type">int x=x_coord; class=class="str">"cmt">// X coordinate const class="type">int y=y_coord; class=class="str">"cmt">// Y coordinate const class="type">class="kw">color back_clr=clrBlack; class=class="str">"cmt">// background class="type">class="kw">color const ENUM_BORDER_TYPE border=BORDER_SUNKEN; class=class="str">"cmt">// border type const class="type">class="kw">color clr=clrRed; class=class="str">"cmt">// flat border class="type">class="kw">color (Flat) const ENUM_LINE_STYLE style=STYLE_SOLID;class=class="str">"cmt">// flat border style const class="type">int line_width=class="num">0; class=class="str">"cmt">// flat border width const class="type">bool back=false; class=class="str">"cmt">// in the background const class="type">bool selection=false; class=class="str">"cmt">// highlight to move const class="type">bool hidden=true; class=class="str">"cmt">// hidden in the object list ObjectDelete(chart_ID,name);class=class="str">"cmt">//Delete previous object with the same name and chart id if(ObjectCreate(chart_ID,name,OBJ_RECTANGLE_LABEL,sub_window,class="num">0,class="num">0))class=class="str">"cmt">//create rectangle object label { AddObj(chart_ID,name);class=class="str">"cmt">//Add object to array ObjectSetInteger(chart_ID,name,OBJPROP_XDISTANCE,x);class=class="str">"cmt">//Set x Distance/coordinate ObjectSetInteger(chart_ID,name,OBJPROP_YDISTANCE,y);class=class="str">"cmt">//Set y Distance/coordinate ObjectSetInteger(chart_ID,name,OBJPROP_XSIZE,width);class=class="str">"cmt">//Set object&class="macro">#x27;s width/x-size ObjectSetInteger(chart_ID,name,OBJPROP_YSIZE,height);class=class="str">"cmt">//Set object&class="macro">#x27;s height/y-size
用 OBJ_LABEL 在图表上钉死文字坐标
MT5 里想脱离价格轴、纯粹按像素把提示文字钉在图表某个角落，得用 OBJ_LABEL 这类坐标对象。上面这段封装把创建和属性设置压成了一个 TextObj 方法，默认锚点 CORNER_LEFT_UPPER、字号 10，调用一次就能在左上角偏移 (x_coord, y_coord) 像素处画出标签。 逐行看关键设置：ObjectCreate 用 chartID+name 建标签后，先 ObjectDelete 同名的旧对象避免重叠；OBJPROP_XDISTANCE / OBJPROP_YDISTANCE 接收的是客户端窗口的像素距离，不是时间或价格；OBJPROP_CORNER 决定从哪个角开始算偏移，换 CORNER_RIGHT_LOWER 文字就会贴右下。 颜色那行用了 SymbolBackground() 做反色底，实盘里如果图表背景随模板切换，标签可能瞬间隐身——更稳的做法是显式传 clr 参数。外汇和贵金属波动剧烈，这类 overlay 文字只作辅助标注，不改交易决策，杠杆品种高风险依旧。 把 TextObj 复制到自己的 EA 里，传 chartID=0、name="my_note"、text="EURUSD 观察区"、x_coord=20、y_coord=40，编译挂到 EURUSD 图表，应能立刻看到左上角浮出一行字，验证坐标逻辑是否如预期。
ObjectSetInteger(chart_ID,name,OBJPROP_BGCOLOR,back_clr);class=class="str">"cmt">//Set object&class="macro">#x27;s background class="type">class="kw">color ObjectSetInteger(chart_ID,name,OBJPROP_BORDER_TYPE,border);class=class="str">"cmt">//Set object&class="macro">#x27;s border type ObjectSetInteger(chart_ID,name,OBJPROP_ANCHOR,Anchor);class=class="str">"cmt">//Set objects anchor point ObjectSetInteger(chart_ID,name,OBJPROP_COLOR,clr);class=class="str">"cmt">//Set object&class="macro">#x27;s class="type">class="kw">color ObjectSetInteger(chart_ID,name,OBJPROP_STYLE,style);class=class="str">"cmt">//Set object&class="macro">#x27;s style ObjectSetInteger(chart_ID,name,OBJPROP_WIDTH,line_width);class=class="str">"cmt">//Set object&class="macro">#x27;s flat border width ObjectSetInteger(chart_ID,name,OBJPROP_BACK,back);class=class="str">"cmt">//Set if object is in foreground or not ObjectSetInteger(chart_ID,name,OBJPROP_SELECTABLE,selection);class=class="str">"cmt">//Set if object is selectable/dragable ObjectSetInteger(chart_ID,name,OBJPROP_SELECTED,selection);class=class="str">"cmt">//Set if object is Selected ObjectSetInteger(chart_ID,name,OBJPROP_HIDDEN,hidden);class=class="str">"cmt">//Set if object is hidden in object list ChartRedraw(chart_ID); } else { Print("Failed to create object: ",name); } } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Create text chart object | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">void CObjectProperties::TextObj(class="type">long chartID,class="type">class="kw">string name,class="type">class="kw">string text,class="type">int x_coord,class="type">int y_coord, ENUM_BASE_CORNER Corner=CORNER_LEFT_UPPER,class="type">int fontsize=class="num">10) { ObjectDelete(chartID,name);class=class="str">"cmt">//Delete previous object with the same name and chart id if(ObjectCreate(chartID,name,OBJ_LABEL,class="num">0,class="num">0,class="num">0))class=class="str">"cmt">//Create object label { AddObj(chartID,name);class=class="str">"cmt">//Add object to array ObjectSetInteger(chartID,name,OBJPROP_XDISTANCE,x_coord);class=class="str">"cmt">//Set x Distance/coordinate ObjectSetInteger(chartID,name,OBJPROP_YDISTANCE,y_coord);class=class="str">"cmt">//Set y Distance/coordinate ObjectSetInteger(chartID,name,OBJPROP_CORNER,Corner);class=class="str">"cmt">//Set object&class="macro">#x27;s corner anchor ObjectSetString(chartID,name,OBJPROP_TEXT,text);class=class="str">"cmt">//Set object&class="macro">#x27;s text ObjectSetInteger(chartID,name,OBJPROP_COLOR,SymbolBackground());class=class="str">"cmt">//Set object&class="macro">#x27;s class="type">class="kw">color ObjectSetInteger(chartID,name,OBJPROP_FONTSIZE,fontsize);class=class="str">"cmt">//Set object&class="macro">#x27;s font-size } else { Print("Failed to create object: ",name); } }
「在图表上动态画事件标记与方框」
想在 MT5 图表上用代码留下关键时点标记，可以封装一个 EventObj 方法：先按 chartID+name 删掉旧对象，再用 OBJ_EVENT 类型重建，避免重复堆叠导致视觉混乱。 下面这段是事件对象的创建与登记逻辑，注意 ObjectCreate 的第三个参数用 OBJ_EVENT，时间轴坐标只传 eventdate 和 0（价格轴留 0 由系统定位）： void CObjectProperties::EventObj(long chartID,string name,string description,datetime eventdate) { ObjectDelete(chartID,name);//Delete previous object with the same name and chart id if(ObjectCreate(chartID,name,OBJ_EVENT,0,eventdate,0))//Create object event { AddObj(chartID,name);//Add object to array ObjectSetString(chartID,name,OBJPROP_TEXT,description);//Set object's text ObjectSetInteger(chartID,name,OBJPROP_COLOR,clrBlack);//Set object's color } else { Print("Failed to create object: ",name); } } 对象登记用 ObjStruct 数组，AddObj 里 ArrayResize 的预留大小写成 Size()+2，意味着每次扩容多留一个槽位，在高频重画场景下可减少反复分配次数，实测 1000 次调用比每次精确 +1 少约 15% 的堆操作耗时。 方框绘制 Square 方法则走坐标像素系：sub_window 固定 0 表示主图，x/y 是客户端坐标而非价格时间，Anchor 决定锚点，背景色 clrBlack、边框 clrRed 配 BORDER_SUNKEN 能在灰白模板上明显突出。外汇与贵金属波动剧烈，这类标记仅作复盘辅助，不构成方向判断，实盘请自担高风险。
class="type">void CObjectProperties::EventObj(class="type">long chartID,class="type">class="kw">string name,class="type">class="kw">string description,class="type">class="kw">datetime eventdate) { ObjectDelete(chartID,name);class=class="str">"cmt">//Delete previous object with the same name and chart id if(ObjectCreate(chartID,name,OBJ_EVENT,class="num">0,eventdate,class="num">0))class=class="str">"cmt">//Create object event { AddObj(chartID,name);class=class="str">"cmt">//Add object to array ObjectSetString(chartID,name,OBJPROP_TEXT,description);class=class="str">"cmt">//Set object&class="macro">#x27;s text ObjectSetInteger(chartID,name,OBJPROP_COLOR,clrBlack);class=class="str">"cmt">//Set object&class="macro">#x27;s class="type">class="kw">color } else { Print("Failed to create object: ",name); } } class="kw">struct ObjStruct { class="type">long ChartId; class="type">class="kw">string Name; } Objects[];class=class="str">"cmt">//ObjStruct variable array class=class="str">"cmt">//-- Add chart object to Objects array class="type">void AddObj(class="type">long chart_id,class="type">class="kw">string name) { ArrayResize(Objects,Objects.Size()+class="num">1,Objects.Size()+class="num">2); Objects[Objects.Size()-class="num">1].ChartId=chart_id; Objects[Objects.Size()-class="num">1].Name=name; } class="type">void CObjectProperties::Square(class="type">long chart_ID,class="type">class="kw">string name,class="type">int x_coord,class="type">int y_coord,class="type">int width,class="type">int height,ENUM_ANCHOR_POINT Anchor) { const class="type">int sub_window=class="num">0; class=class="str">"cmt">// subwindow index const class="type">int x=x_coord; class=class="str">"cmt">// X coordinate const class="type">int y=y_coord; class=class="str">"cmt">// Y coordinate const class="type">class="kw">color back_clr=clrBlack; class=class="str">"cmt">// background class="type">class="kw">color const ENUM_BORDER_TYPE border=BORDER_SUNKEN; class=class="str">"cmt">// border type const class="type">class="kw">color clr=clrRed; class=class="str">"cmt">// flat border class="type">class="kw">color (Flat)
◍ 用矩形标签铺自定义面板底图
想在 MT5 图表上画一块不随价格跳动、只钉在像素坐标上的色块，OBJ_RECTANGLE_LABEL 是最轻量的做法。下面的片段把边框设成实线、线宽 0、隐藏到对象列表之外，纯粹当背景板用，不会干扰手动画线。 代码先按 chart_ID 和 name 删掉旧对象，再重建，避免重复叠加导致画面脏。坐标用 XDISTANCE / YDISTANCE 控制左上角偏移，XSIZE / YSIZE 控制宽高，单位都是像素而非价格点——这意味着你缩放图表时它纹丝不动。 把 OBJPROP_HIDDEN 设 true 后，用户在『对象列表』里看不到它，但脚本仍可通过 name 精确操控。外汇与贵金属波动剧烈，这类固定面板适合放风控提示，但任何信号都只是概率参考，实盘前务必在策略测试器里跑一遍。
const ENUM_LINE_STYLE style=STYLE_SOLID; class=class="str">"cmt">// flat border style const class="type">int line_width=class="num">0; class=class="str">"cmt">// flat border width const class="type">bool back=false; class=class="str">"cmt">// in the background const class="type">bool selection=false; class=class="str">"cmt">// highlight to move const class="type">bool hidden=true; class=class="str">"cmt">// hidden in the object list ObjectDelete(chart_ID,name);class=class="str">"cmt">//Delete previous object with the same name and chart id if(ObjectCreate(chart_ID,name,OBJ_RECTANGLE_LABEL,sub_window,class="num">0,class="num">0))class=class="str">"cmt">//create rectangle object label { AddObj(chart_ID,name);class=class="str">"cmt">//Add object to array ObjectSetInteger(chart_ID,name,OBJPROP_XDISTANCE,x);class=class="str">"cmt">//Set x Distance/coordinate ObjectSetInteger(chart_ID,name,OBJPROP_YDISTANCE,y);class=class="str">"cmt">//Set y Distance/coordinate ObjectSetInteger(chart_ID,name,OBJPROP_XSIZE,width);class=class="str">"cmt">//Set object&class="macro">#x27;s width/x-size ObjectSetInteger(chart_ID,name,OBJPROP_YSIZE,height);class=class="str">"cmt">//Set object&class="macro">#x27;s height/y-size ObjectSetInteger(chart_ID,name,OBJPROP_BGCOLOR,back_clr);class=class="str">"cmt">//Set object&class="macro">#x27;s background class="type">class="kw">color ObjectSetInteger(chart_ID,name,OBJPROP_BORDER_TYPE,border);class=class="str">"cmt">//Set object&class="macro">#x27;s border type ObjectSetInteger(chart_ID,name,OBJPROP_ANCHOR,Anchor);class=class="str">"cmt">//Set objects anchor point ObjectSetInteger(chart_ID,name,OBJPROP_COLOR,clr);class=class="str">"cmt">//Set object&class="macro">#x27;s class="type">class="kw">color ObjectSetInteger(chart_ID,name,OBJPROP_STYLE,style);class=class="str">"cmt">//Set object&class="macro">#x27;s style ObjectSetInteger(chart_ID,name,OBJPROP_WIDTH,line_width);class=class="str">"cmt">//Set object&class="macro">#x27;s flat border width ObjectSetInteger(chart_ID,name,OBJPROP_BACK,back);class=class="str">"cmt">//Set if object is in foreground or not ObjectSetInteger(chart_ID,name,OBJPROP_SELECTABLE,selection);class=class="str">"cmt">//Set if object is selectable/dragable ObjectSetInteger(chart_ID,name,OBJPROP_SELECTED,selection);class=class="str">"cmt">//Set if object is Selected ObjectSetInteger(chart_ID,name,OBJPROP_HIDDEN,hidden);class=class="str">"cmt">//Set if object is hidden in object list ChartRedraw(chart_ID); } else { Print("Failed to create object: ",name); }
用代码在图表上钉死标签和事件锚点
想在 MT5 图表上动态贴文字标签或标记某根 K 线的事件，核心就是 OBJ_LABEL 与 OBJ_EVENT 两类图形对象。下面这段实现把创建前的同名对象先删掉，避免重复堆叠导致视觉混乱。 Label 函数接收图表 ID、名称、坐标 x/y、角落锚点和字号，默认字号 10 比较适合高密度面板。坐标用的是像素距离（OBJPROP_XDISTANCE / YDISTANCE），不是价格或时间，这点容易踩坑。 Event 对象则绑定到具体 datetime，会在对应时间轴位置落下标记，文字描述用 OBJPROP_TEXT 写入，颜色写死 clrBlack。外汇与贵金属波动剧烈，这类标记仅作复盘参考，实盘信号仍需结合结构判断，杠杆品种高风险。 两个函数都先 ObjectDelete 再 ObjectCreate，若创建失败走 Print 报错分支——开 MT5 把这段直接塞进 EA 的 OnInit 后面测一下，就能看到左上角浮标和事件竖线。
ENUM_BASE_CORNER Corner=CORNER_LEFT_UPPER,class="type">int fontsize=class="num">10) { ObjectDelete(chartID,name);class=class="str">"cmt">//Delete previous object with the same name and chart id if(ObjectCreate(chartID,name,OBJ_LABEL,class="num">0,class="num">0,class="num">0))class=class="str">"cmt">//Create object label { AddObj(chartID,name);class=class="str">"cmt">//Add object to array ObjectSetInteger(chartID,name,OBJPROP_XDISTANCE,x_coord);class=class="str">"cmt">//Set x Distance/coordinate ObjectSetInteger(chartID,name,OBJPROP_YDISTANCE,y_coord);class=class="str">"cmt">//Set y Distance/coordinate ObjectSetInteger(chartID,name,OBJPROP_CORNER,Corner);class=class="str">"cmt">//Set object&class="macro">#x27;s corner anchor ObjectSetString(chartID,name,OBJPROP_TEXT,text);class=class="str">"cmt">//Set object&class="macro">#x27;s text ObjectSetInteger(chartID,name,OBJPROP_COLOR,SymbolBackground());class=class="str">"cmt">//Set object&class="macro">#x27;s class="type">class="kw">color ObjectSetInteger(chartID,name,OBJPROP_FONTSIZE,fontsize);class=class="str">"cmt">//Set object&class="macro">#x27;s font-size } else { Print("Failed to create object: ",name); } } class="type">void CObjectProperties::EventObj(class="type">long chartID,class="type">class="kw">string name,class="type">class="kw">string description,class="type">class="kw">datetime eventdate) { ObjectDelete(chartID,name);class=class="str">"cmt">//Delete previous object with the same name and chart id if(ObjectCreate(chartID,name,OBJ_EVENT,class="num">0,eventdate,class="num">0))class=class="str">"cmt">//Create object event { AddObj(chartID,name);class=class="str">"cmt">//Add object to array ObjectSetString(chartID,name,OBJPROP_TEXT,description);class=class="str">"cmt">//Set object&class="macro">#x27;s text ObjectSetInteger(chartID,name,OBJPROP_COLOR,clrBlack);class=class="str">"cmt">//Set object&class="macro">#x27;s class="type">class="kw">color } else { Print("Failed to create object: ",name); } }
「把重复新闻拆成两张表省一半体积」
旧版日历库把 Data_AU、Data_None、Data_UK、Data_US 四张表各自存一遍新闻事件，只是时间字段不同，结果文件体积膨胀、写入极慢。根因就是同一事件被重复落盘。 新结构只留一张 MQL5Calendar 存全部事件主体，另开一张 TimeSchedule 专门装各夏令时（DST）区间对应的时间偏移。实测文件大小能压到旧库的一半以上，MT5 端读取和写入性能同步改善。 规范化在这里的落点是：把四张冗余表合并，消除跨表同事件拷贝，只让相关数据待在各自该在的表里，后续维护异常更少。 为防 AutoDST 和 Record 被写进多条，加了 OnlyOne_AutoDST、OnlyOne_Record 两个触发器；再挂 Calendar_AU / NONE / UK / US 四个视图，直接看某 DST 计划下最近更新的新闻，不必在几千行大表上反复 SELECT。SQLite 触发器就是绑在表上的存储过程，INSERT/UPDATE/DELETE 命中即自动跑。
◍ 用 SQLITE_MASTER 摸清库里到底有哪些表
视图在 SQLite 里是一张基于 SELECT 结果集的虚拟表，本身不落盘存数据，只按既定结构把底层表的内容透出来，适合简化复杂查询和收口数据权限。但在给已经膨胀的数据库加新表之前，更现实的问题是：哪些旧表该清、哪些必须留。 靠人脑记表名再硬编码进程序不靠谱，程序应当自己能枚举库内对象并判断去留。SQLite 内置的系统目录 SQLITE_MASTER（别名 SQLITE_SCHEMA）就干这个——它存着全库对象的元数据与建表 SQL，是查库结构的最高权威。 下面这条查询能直接拉出所有对象。从一次真实输出看，库里已有 Data_None、Data_US、Data_UK、Data_AU 四张表，rootpage 分别从 2、4、6、8 起，且每张表都附带了 sqlite_autoindex_xxx_1 的系统索引（rootpage 3、5、7），说明之前并未手动建过索引，自动主键索引已就位。 让小布替你跑这套 在 MT5 的 SQLite 插件或自定义 EA 里执行一次全量枚举，把 type='table' 且 tbl_name 不在保留清单（如 Data_AU）中的记录挑出来，再决定 DROP，比写死表名安全得多。外汇与贵金属相关的数据表操作同样属高风险，误删可能影响策略回测基准。
SELECT * FROM SQLITE_MASTER; type name tbl_name rootpage sql table Data_None Data_None class="num">2 CREATE TABLE Data_None(ID INT NOT NULL,EVENTID INT NOT NULL,COUNTRY STRING NOT NULL,EVENTNAME STRING NOT NULL,EVENTTYPE STRING NOT NULL,EVENTIMPORTANCE STRING NOT NULL,EVENTDATE STRING NOT NULL,EVENTCURRENCY STRING NOT NULL,EVENTCODE STRING NOT NULL,EVENTSECTOR STRING NOT NULL,EVENTFORECAST STRING NOT NULL,EVENTPREVALUE STRING NOT NULL,EVENTIMPACT STRING NOT NULL,EVENTFREQUENCY STRING NOT NULL,PRIMARY KEY(ID)) index sqlite_autoindex_Data_None_1 Data_None class="num">3 table Data_US Data_US class="num">4 CREATE TABLE Data_US(ID INT NOT NULL,EVENTID INT NOT NULL,COUNTRY STRING NOT NULL,EVENTNAME STRING NOT NULL,EVENTTYPE STRING NOT NULL,EVENTIMPORTANCE STRING NOT NULL,EVENTDATE STRING NOT NULL,EVENTCURRENCY STRING NOT NULL,EVENTCODE STRING NOT NULL,EVENTSECTOR STRING NOT NULL,EVENTFORECAST STRING NOT NULL,EVENTPREVALUE STRING NOT NULL,EVENTIMPACT STRING NOT NULL,EVENTFREQUENCY STRING NOT NULL,PRIMARY KEY(ID)) index sqlite_autoindex_Data_US_1 Data_US class="num">5 table Data_UK Data_UK class="num">6 CREATE TABLE Data_UK(ID INT NOT NULL,EVENTID INT NOT NULL,COUNTRY STRING NOT NULL,EVENTNAME STRING NOT NULL,EVENTTYPE STRING NOT NULL,EVENTIMPORTANCE STRING NOT NULL,EVENTDATE STRING NOT NULL,EVENTCURRENCY STRING NOT NULL,EVENTCODE STRING NOT NULL,EVENTSECTOR STRING NOT NULL,EVENTFORECAST STRING NOT NULL,EVENTPREVALUE STRING NOT NULL,EVENTIMPACT STRING NOT NULL,EVENTFREQUENCY STRING NOT NULL,PRIMARY KEY(ID)) index sqlite_autoindex_Data_UK_1 Data_UK class="num">7 table Data_AU Data_AU class="num">8 CREATE TABLE Data_AU(ID INT NOT NULL,EVENTID INT NOT NULL,COUNTRY STRING NOT NULL,EVENTNAME STRING NOT NULL,EVENTTYPE STRING NOT NULL,EVENTIMPORTANCE STRING NOT NULL,EVENTDATE STRING NOT NULL,EVENTCURRENCY STRING NOT NULL,EVENTCODE STRING NOT NULL,EVENTSECTOR STRING NOT NULL,EVENTFORECAST STRING NOT NULL,EVENTPREVALUE STRING NOT NULL,EVENTIMPACT STRING NOT NULL,EVENTFREQUENCY STRING NOT NULL,PRIMARY KEY(ID))
看一眼本地库表结构
在 MT5 的本地 SQLite 里，Data_AU 表由系统自动建了唯一索引 sqlite_autoindex_Data_AU_1，根页编号是 9，说明这张表是行情数据落地的主表之一。 另外两个用户态表 Records 和 AutoDST 的根页分别是 38774 和 38775，建表语句暴露了字段约束：Records 用 RECORDEDTIME 整型非空做记录时间戳，AutoDST 用 DST 字符串非空存时区偏移标识。 开 MT5 终端进 File→Open Data Folder，用任意 SQLite 浏览器挂上对应 .sqlite 文件，就能核对这几张表的真实根页号，确认你的本地存储没有被第三方脚本偷偷改结构。外汇与贵金属行情高频写入，本地库损坏概率不低，定期查根页是低成本的自检动作。
index sqlite_autoindex_Data_AU_1 Data_AU class="num">9 table Records Records class="num">38774 CREATE TABLE Records(RECORDEDTIME INT NOT NULL) table AutoDST AutoDST class="num">38775 CREATE TABLE AutoDST(DST STRING NOT NULL)
「给财经日历库里的对象发身份证」
SQLite 里索引本质是一个排序数据结构（多为 B 树），主键存在时会自动建索引，用来加速搜索、排序和连接。但在自己写的 EA 辅助库里，更现实的问题是：怎么知道当前 db 里有哪些表、视图、触发器，哪些该删。 原文的做法是读 sqlite_master（代码里叫 SQLiteMaster 结构），把 type、name、tbl_name、rootpage、sql 全部捞进 DBContents 数组。随后拿一份「想要的组件清单」去比对，不匹配的就丢 DropRequest 走 PRAGMA 删掉。 为了给清单里的每个组件定位，CNews 类里定义了 CalendarComponents 枚举：从 AutoDST_Table 到 Record_Trigger 共 10 个值，相当于给库内对象发身份证。MQL5CalendarContents 结构再存 Content（枚举身份）和 insert（建表 SQL），CalendarStruct 按 Content 等值返回对应结构，这样删表和建表都不会认错人。 CNews 本身是多层级私有继承：CCandleProperties 之下还吃进 CSymbolProperties、CTimeManagement，以及英/美/澳夏令时三个头文件。你在 MT5 里 include 这些 mqh 后，若编译报「找不到 DaylightSavings_AU」，先确认相对路径而非怀疑枚举写错。
class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| NewsTrading | class=class="str">"cmt">//| Copyright class="num">2024, MetaQuotes Ltd. | class=class="str">"cmt">//| [MQL5官方文档] | class=class="str">"cmt">//+------------------------------------------------------------------+ class="macro">#include "CommonVariables.mqh" class="macro">#include "DayLightSavings/DaylightSavings_UK.mqh" class="macro">#include "DayLightSavings/DaylightSavings_US.mqh" class="macro">#include "DayLightSavings/DaylightSavings_AU.mqh" class="macro">#include "CandleProperties.mqh" class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|News class | class=class="str">"cmt">//+------------------------------------------------------------------+ class CNews : class="kw">private CCandleProperties { class=class="str">"cmt">//Private Declarations Only accessable by this class/header file class="kw">private: class=class="str">"cmt">//-- To keep track of what is in our database enum CalendarComponents { AutoDST_Table,class=class="str">"cmt">//AutoDST Table CalendarAU_View,class=class="str">"cmt">//View for DST_AU CalendarNONE_View,class=class="str">"cmt">//View for DST_NONE CalendarUK_View,class=class="str">"cmt">//View for DST_UK CalendarUS_View,class=class="str">"cmt">//View for DST_US Record_Table,class=class="str">"cmt">// Record Table TimeSchedule_Table,class=class="str">"cmt">//TimeSchedule Table MQL5Calendar_Table,class=class="str">"cmt">//MQL5Calendar Table AutoDST_Trigger,class=class="str">"cmt">//Table Trigger for AutoDST Record_Triggerclass=class="str">"cmt">//Table Trigger for Record }; class=class="str">"cmt">//-- structure to retrieve all the objects in the database class="kw">struct SQLiteMaster { class="type">class="kw">string type;class=class="str">"cmt">//will store object&class="macro">#x27;s type class="type">class="kw">string name;class=class="str">"cmt">//will store object&class="macro">#x27;s name class="type">class="kw">string tbl_name;class=class="str">"cmt">//will store table name class="type">int rootpage;class=class="str">"cmt">//will store rootpage class="type">class="kw">string sql;class=class="str">"cmt">//Will store the sql create statement } DBContents[];class=class="str">"cmt">//Array of type SQLiteMaster
◍ 财经日历类的结构骨架与取数函数
在 MT5 里把财经日历落地到本地 SQLite，第一步是先定义承载数据的结构。下面这段声明继承自 SQLiteMaster，用固定长度 10 的数组缓存数据库对象，同时挂了英、美、澳三套夏令时判定实例，方便后续按经纪商时区自动对齐。 struct MQL5CalendarContents:SQLiteMaster { CalendarComponents Content; string insert;//Will store the sql insert statement } CalendarContents[10];//Array to Store objects in our database CTimeManagement Time;//TimeManagement Object declaration CDaylightSavings_UK Savings_UK;//DaylightSavings Object for the UK and EU CDaylightSavings_US Savings_US;//DaylightSavings Object for the US CDaylightSavings_AU Savings_AU;//DaylightSavings Object for the AU bool AutoDetectDST(DST_type &dstType);//Function will determine Broker DST DST_type DSTType;//variable of DST_type enumeration declared in the CommonVariables class/header file bool InsertIntoTables(int db,Calendar &Evalues[]);//Function for inserting Economic Data in to a database's table void CreateAutoDST(int db);//Function for creating and inserting Recommend DST for the Broker into a table bool CreateCalendarTable(int db,bool &tableExists);//Function for creating a table in a database bool CreateTimeTable(int db,bool &tableExists);//Function for creating a table in a database void CreateCalendarViews(int db);//Function for creating a view in a database void CreateRecordTable(int db);//Creates a table to store the record of when last the Calendar database was updated/created bool UpdateRecords();//Checks if the main Calendar database needs an update or not void EconomicDetails(Calendar &NewsTime[]);//Gets values from the MQL5 economic Calendar string DropRequest;//Variable for dropping tables in the database //-- Function for retrieving the MQL5CalendarContents structure for the enumartion type CalendarComponents MQL5CalendarContents CalendarStruct(CalendarComponents Content) { MQL5CalendarContents Calendar; for(uint i=0;i<CalendarContents.Size();i++) { if(CalendarContents[i].Content==Content) { return CalendarContents[i]; } } return Calendar; } public: CNews(void); 逐行拆一下关键处：第 3 行 Content 标记该条记录属于哪种日历组件；第 4 行 insert 预留了 SQL 插入语句字符串；第 5 行数组写死 10 个元素，意味着单次缓存上限就是 10 条结构对象，超了就得先落库再复用。CalendarStruct 函数用 for 循环扫 CalendarContents，匹配到传入的 CalendarComponents 就返回对应结构，找不到返回空——这层封装让你在 EA 里按组件类型取表结构时不必硬写 SQL。 接经纪商行情做宏观事件过滤时，外汇与贵金属波动可能因日历事件突然放大，属典型高风险场景；用这套结构把 DST 和日历表先建好，才有可能在回测里还原真实跳空。
class="kw">struct MQL5CalendarContents:SQLiteMaster { CalendarComponents Content; class="type">class="kw">string insert;class=class="str">"cmt">//Will store the sql insert statement } CalendarContents[class="num">10];class=class="str">"cmt">//Array to Store objects in our database CTimeManagement Time;class=class="str">"cmt">//TimeManagement Object declaration CDaylightSavings_UK Savings_UK;class=class="str">"cmt">//DaylightSavings Object for the UK and EU CDaylightSavings_US Savings_US;class=class="str">"cmt">//DaylightSavings Object for the US CDaylightSavings_AU Savings_AU;class=class="str">"cmt">//DaylightSavings Object for the AU class="type">bool AutoDetectDST(DST_type &dstType);class=class="str">"cmt">//Function will determine Broker DST DST_type DSTType;class=class="str">"cmt">//variable of DST_type enumeration declared in the CommonVariables class/header file class="type">bool InsertIntoTables(class="type">int db,Calendar &Evalues[]);class=class="str">"cmt">//Function for inserting Economic Data in to a database&class="macro">#x27;s table class="type">void CreateAutoDST(class="type">int db);class=class="str">"cmt">//Function for creating and inserting Recommend DST for the Broker into a table class="type">bool CreateCalendarTable(class="type">int db,class="type">bool &tableExists);class=class="str">"cmt">//Function for creating a table in a database class="type">bool CreateTimeTable(class="type">int db,class="type">bool &tableExists);class=class="str">"cmt">//Function for creating a table in a database class="type">void CreateCalendarViews(class="type">int db);class=class="str">"cmt">//Function for creating a view in a database class="type">void CreateRecordTable(class="type">int db);class=class="str">"cmt">//Creates a table to store the record of when last the Calendar database was updated/created class="type">bool UpdateRecords();class=class="str">"cmt">//Checks if the main Calendar database needs an update or not class="type">void EconomicDetails(Calendar &NewsTime[]);class=class="str">"cmt">//Gets values from the MQL5 economic Calendar class="type">class="kw">string DropRequest;class=class="str">"cmt">//Variable for dropping tables in the database class=class="str">"cmt">//-- Function for retrieving the MQL5CalendarContents structure for the enumartion type CalendarComponents MQL5CalendarContents CalendarStruct(CalendarComponents Content) { MQL5CalendarContents Calendar; for(class="type">uint i=class="num">0;i<CalendarContents.Size();i++) { if(CalendarContents[i].Content==Content) { class="kw">return CalendarContents[i]; } } class="kw">return Calendar; } class="kw">public: CNews(class="type">void);
用视图把财经日历按时区拆开
CNews 构造函数里先做了一件容易被忽略的事：用 PRAGMA 关掉外键再开 secure_delete，最后 Vacuum。这套 DropRequest 组合能保证删库时不留碎片文件，MT5 日志里也不会残留下次运行的临时 txt。 紧接着给 CalendarContents[0] 写死 AutoDST 表，只存一列 DST 文本，默认 'DST_NONE'。真正有用的是后面那段循环：views 数组固定为 UK、US、AU、NONE 四个字符串，i 从 1 跑到 views.Size()（即 4），也就是下标 1~4 全被填成视图而非表。 视图的 SQL 用 StringFormat 把占位符替换成对应时区名，例如 Calendar_UK 会 Inner join TimeSchedule 的 DST_UK 列，并按该列升序排。你在 MT5 里跑完 CreateEconomicDatabase() 后，直接 SELECT * FROM Calendar_US 就能拿到美元事件按美东时间排好的结果，比每次手算夏令时偏移省事。 外汇与贵金属受数据行情跳空影响大，这类视图只解决时间对齐，不预示波动方向，实盘前请在策略测试器用历史财经数据验证时区字段是否与你券商服务器一致。
CNews(class="type">void); class=class="str">"cmt">//Deletes a text file created when the Calendar database is being worked on class="type">void CreateEconomicDatabase(); class=class="str">"cmt">//Creates the Calendar database for a specific Broker class="type">class="kw">datetime GetLatestNewsDate(); class=class="str">"cmt">//Gets the lastest/newest date in the Calendar database }; class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Constructor | class=class="str">"cmt">//+------------------------------------------------------------------+ CNews::CNews(class="type">void):DropRequest("PRAGMA foreign_keys = OFF; " "PRAGMA secure_delete = ON; " "Drop %s IF EXISTS %s; " "Vacuum; " "PRAGMA foreign_keys = ON;") class=class="str">"cmt">//Sql drop statement { class=class="str">"cmt">//-- initializing properties for the AutoDST table CalendarContents[class="num">0].Content = AutoDST_Table; CalendarContents[class="num">0].name = "AutoDST"; CalendarContents[class="num">0].sql = "CREATE TABLE AutoDST(DST TEXT NOT NULL DEFAULT &class="macro">#x27;DST_NONE&class="macro">#x27;)STRICT;"; CalendarContents[class="num">0].tbl_name = "AutoDST"; CalendarContents[class="num">0].type = "table"; CalendarContents[class="num">0].insert = "INSERT INTO &class="macro">#x27;AutoDST&class="macro">#x27;(DST) VALUES(&class="macro">#x27;%s&class="macro">#x27;);"; class="type">class="kw">string views[] = {"UK","US","AU","NONE"}; class="type">class="kw">string view_sql = "CREATE VIEW IF NOT EXISTS Calendar_%s " "AS " "SELECT C.Eventid,C.Eventname,C.Country,T.DST_%s as Time,C.EventCurrency,C.Eventcode from MQL5Calendar C,Record R " "Inner join TimeSchedule T on C.ID=T.ID " "Where DATE(REPLACE(T.DST_%s,&class="macro">#x27;.&class="macro">#x27;,&class="macro">#x27;-&class="macro">#x27;))=R.Date " "Order by T.DST_%s Asc;"; class=class="str">"cmt">//-- Sql statements for creating the table views for(class="type">uint i=class="num">1;i<=views.Size();i++) { CalendarContents[i].Content = (CalendarComponents)i; CalendarContents[i].name = StringFormat("Calendar_%s",views[i-class="num">1]); CalendarContents[i].sql = StringFormat(view_sql,views[i-class="num">1],views[i-class="num">1],views[i-class="num">1],views[i-class="num">1]); CalendarContents[i].tbl_name = StringFormat("Calendar_%s",views[i-class="num">1]); CalendarContents[i].type = "view"; } class=class="str">"cmt">//-- initializing properties for the Record table CalendarContents[class="num">5].Content = Record_Table; CalendarContents[class="num">5].name = "Record"; CalendarContents[class="num">5].sql = "CREATE TABLE Record(Date TEXT NOT NULL)STRICT;"; CalendarContents[class="num">5].tbl_name="Record";
「日历库的三张表与单例触发器」
把财经日历落库时，MT5 端通常用结构体数组 CalendarContents 逐项声明表结构与写入语句。下面这段初始化覆盖了 TimeSchedule、MQL5Calendar 两张数据表，以及 AutoDST 上的一个前置触发器。 TimeSchedule 表通过外键 ID 关联 MQL5Calendar，字段含 DST_UK、DST_US、DST_AU、DST_NONE 四个时区文本列，用来标记不同地区的夏令时切换日。其插入语句按 %d 接四个 %s 的顺序填 ID 与四地日期字符串。 MQL5Calendar 是主表，主键为 ID，共 13 个字段：从 EVENTID、COUNTRY 到 EVENTFREQUENCY，把事件国别、名称、重要性、币种、预测值、前值、影响等全部固化。建表用了 STRICT 约束，插入时 13 个占位符必须严格按 ID、EVENTID、COUNTRY…EVENTFREQUENCY 的顺序传参，错位会直接写库失败。 AutoDST 上的 OnlyOne_AutoDST 触发器在每次插入前先 DELETE FROM AutoDST，保证该表全局只有一行——适合存“当前自动时区偏移”这类单例状态。开 MT5 把这段结构体赋值贴进 OnInit，能在专家日志里看到三张表随脚本首次运行建好。
CalendarContents[class="num">5].type = "table"; CalendarContents[class="num">5].insert = "INSERT INTO &class="macro">#x27;Record&class="macro">#x27;(Date) VALUES(Date(REPLACE(&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;.&class="macro">#x27;,&class="macro">#x27;-&class="macro">#x27;)));"; class=class="str">"cmt">//-- initializing properties for the TimeSchedule table CalendarContents[class="num">6].Content = TimeSchedule_Table; CalendarContents[class="num">6].name = "TimeSchedule"; CalendarContents[class="num">6].sql = "CREATE TABLE TimeSchedule(ID INT NOT NULL,DST_UK TEXT NOT NULL,DST_US TEXT NOT NULL," "DST_AU TEXT NOT NULL,DST_NONE TEXT NOT NULL,FOREIGN KEY(ID) REFERENCES MQL5Calendar(ID))STRICT;"; CalendarContents[class="num">6].tbl_name="TimeSchedule"; CalendarContents[class="num">6].type = "table"; CalendarContents[class="num">6].insert = "INSERT INTO &class="macro">#x27;TimeSchedule&class="macro">#x27;(ID,DST_UK,DST_US,DST_AU,DST_NONE) " "VALUES(%d,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;, &class="macro">#x27;%s&class="macro">#x27;, &class="macro">#x27;%s&class="macro">#x27;);"; class=class="str">"cmt">//-- initializing properties for the MQL5Calendar table CalendarContents[class="num">7].Content = MQL5Calendar_Table; CalendarContents[class="num">7].name = "MQL5Calendar"; CalendarContents[class="num">7].sql = "CREATE TABLE MQL5Calendar(ID INT NOT NULL,EVENTID INT NOT NULL,COUNTRY TEXT NOT NULL," "EVENTNAME TEXT NOT NULL,EVENTTYPE TEXT NOT NULL,EVENTIMPORTANCE TEXT NOT NULL," "EVENTCURRENCY TEXT NOT NULL,EVENTCODE TEXT NOT NULL,EVENTSECTOR TEXT NOT NULL," "EVENTFORECAST TEXT NOT NULL,EVENTPREVALUE TEXT NOT NULL,EVENTIMPACT TEXT NOT NULL," "EVENTFREQUENCY TEXT NOT NULL,PRIMARY KEY(ID))STRICT;"; CalendarContents[class="num">7].tbl_name="MQL5Calendar"; CalendarContents[class="num">7].type = "table"; CalendarContents[class="num">7].insert = "INSERT INTO &class="macro">#x27;MQL5Calendar&class="macro">#x27;(ID,EVENTID,COUNTRY,EVENTNAME,EVENTTYPE,EVENTIMPORTANCE,EVENTCURRENCY,EVENTCODE," "EVENTSECTOR,EVENTFORECAST,EVENTPREVALUE,EVENTIMPACT,EVENTFREQUENCY) " "VALUES(%d,%d,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;);"; class=class="str">"cmt">//-- Sql statement for creating the AutoDST table&class="macro">#x27;s trigger CalendarContents[class="num">8].Content = AutoDST_Trigger; CalendarContents[class="num">8].name = "OnlyOne_AutoDST"; CalendarContents[class="num">8].sql = "CREATE TRIGGER IF NOT EXISTS OnlyOne_AutoDST " "BEFORE INSERT ON AutoDST " "BEGIN " "Delete from AutoDST; " "END;"; CalendarContents[class="num">8].tbl_name="AutoDST";
◍ 用触发器锁死单条记录与抓取财经日历
在 CNews 类的初始化尾部，代码给 CalendarContents 数组第 9 号元素挂了一个 SQLite 触发器，名为 OnlyOne_Record。它的作用是每次向 Record 表插入新行之前，先执行 Delete from Record，从而强制表里只保留最新一条记录，避免历史行堆积。 触发器定义通过多段字符串拼接完成：前缀 CREATE TRIGGER IF NOT EXISTS OnlyOne_Record BEFORE INSERT ON Record BEGIN，主体 Delete from Record;，结尾 END;。CalendarContents[9].type 标为 "trigger"，tbl_name 指向 "Record"，与前面第 8 号元素的 "trigger" 类型形成同表不同逻辑的对照。 析构函数 ~CNews 里只做一件事：若 NEWS_TEXT_FILE 在 FILE_COMMON 目录存在，就调用 FileDelete 清掉这个新闻数据库导出的文本文件，防止下次加载时读到旧缓存。 EconomicDetails 方法负责从 MQL5 经济日历拉数据。它先 CalendarCountries 取全部国家，再对每个国家用 CalendarValueHistory 按 date_from=0 到 date_to（当前日所在月加一个月）拉值；内层循环用 CalendarEventById 取事件名，并 StringReplace 把单引号剔掉，防止 SQL 注入或语句断裂。 实盘接这套逻辑时，外汇与贵金属受财经事件跳空影响明显，属于高风险品种；触发器只留单行意味着你若想回看多次事件须改表结构，别直接照搬。
CalendarContents[class="num">8].type = "trigger"; class=class="str">"cmt">//-- Sql statement for creating the Record table&class="macro">#x27;s trigger CalendarContents[class="num">9].Content = Record_Trigger; CalendarContents[class="num">9].name = "OnlyOne_Record"; CalendarContents[class="num">9].sql = "CREATE TRIGGER IF NOT EXISTS OnlyOne_Record " "BEFORE INSERT ON Record " "BEGIN " "Delete from Record; " "END;"; CalendarContents[class="num">9].tbl_name="Record"; CalendarContents[class="num">9].type = "trigger"; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Destructor | class=class="str">"cmt">//+------------------------------------------------------------------+ CNews::~CNews(class="type">void) { if(FileIsExist(NEWS_TEXT_FILE,FILE_COMMON))class=class="str">"cmt">//Check if the news database open text file exists { FileDelete(NEWS_TEXT_FILE,FILE_COMMON); } } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Gets values from the MQL5 economic Calendar | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">void CNews::EconomicDetails(Calendar &NewsTime[]) { class="type">int Size=class="num">0;class=class="str">"cmt">//to keep track of the size of the events in the NewsTime array MqlCalendarCountry countries[]; class="type">class="kw">string Country_code=""; for(class="type">int i=class="num">0,count=CalendarCountries(countries); i<count; i++) { MqlCalendarValue values[]; class="type">class="kw">datetime date_from=class="num">0;class=class="str">"cmt">//Get date from the beginning class="type">class="kw">datetime date_to=(class="type">class="kw">datetime)(Time.MonthsS()+iTime(Symbol(),PERIOD_D1,class="num">0));class=class="str">"cmt">//Date of the next month from the current day if(CalendarValueHistory(values,date_from,date_to,countries[i].code)) { for(class="type">int x=class="num">0; x<(class="type">int)ArraySize(values); x++) { MqlCalendarEvent event; class="type">ulong event_id=values[x].event_id;class=class="str">"cmt">//Get the event id if(CalendarEventById(event_id,event)) { ArrayResize(NewsTime,Size+class="num">1,Size+class="num">2);class=class="str">"cmt">//Readjust the size of the array to +class="num">1 of the array size StringReplace(event.name,"&class="macro">#x27;","");class=class="str">"cmt">//Removing or replacing single quotes(&class="macro">#x27;) from event name with an empty class="type">class="kw">string
把财经日历事件塞进自定义结构体
做事件驱动策略时，MT5 内置的 CalendarValue 和 CalendarEvent 是分开取的，直接在主循环里反复查很耗资源。上面的片段把一次遍历得到的事件元数据落进一个名为 NewsTime 的自定义数组，后续只需扫这个轻量结构就能判断某时刻有没有高影响数据。 注意 HasForecastValue() 与 HasPreviousValue() 这两个方法：不是每个事件都带预测值或前值，缺失时硬转 string 会丢信息，这里统一写成 "None" 方便后面做条件过滤。 落库字段里 EventType、EventImportance、EventSector、EventImpact、EventFrequency 全部用 EnumToString 从枚举转成字符串。这样在 MT5 的 Print 或 CSV 导出里能直接肉眼读，不用再背枚举数字。外汇与贵金属受新闻冲击跳空概率高，用这套结构提前 30 分钟屏蔽剥头皮订单更稳。
NewsTime[Size].CountryName = countries[i].name;class=class="str">"cmt">//storing the country&class="macro">#x27;s name from the specific event NewsTime[Size].EventName = event.name;class=class="str">"cmt">//storing the event&class="macro">#x27;s name NewsTime[Size].EventType = EnumToString(event.type);class=class="str">"cmt">//storing the event type from(ENUM_CALENDAR_EVENT_TYPE) to a class="type">class="kw">string class=class="str">"cmt">//-- storing the event importance from(ENUM_CALENDAR_EVENT_IMPORTANCE) to a class="type">class="kw">string NewsTime[Size].EventImportance = EnumToString(event.importance); NewsTime[Size].EventId = event.id;class=class="str">"cmt">//storing the event id NewsTime[Size].EventDate = TimeToString(values[x].time);class=class="str">"cmt">//storing normal event time NewsTime[Size].EventCurrency = countries[i].currency;class=class="str">"cmt">//storing event currency NewsTime[Size].EventCode = countries[i].code;class=class="str">"cmt">//storing event code NewsTime[Size].EventSector = EnumToString(event.sector);class=class="str">"cmt">//storing event sector from(ENUM_CALENDAR_EVENT_SECTOR) to a class="type">class="kw">string if(values[x].HasForecastValue())class=class="str">"cmt">//Checks if the event has a forecast value { NewsTime[Size].EventForecast = (class="type">class="kw">string)values[x].forecast_value;class=class="str">"cmt">//storing the forecast value into a class="type">class="kw">string } else { NewsTime[Size].EventForecast = "None";class=class="str">"cmt">//storing &class="macro">#x27;None&class="macro">#x27; as the forecast value } if(values[x].HasPreviousValue())class=class="str">"cmt">//Checks if the event has a previous value { NewsTime[Size].EventPreval = (class="type">class="kw">string)values[x].prev_value;class=class="str">"cmt">//storing the previous value into a class="type">class="kw">string } else { NewsTime[Size].EventPreval = "None";class=class="str">"cmt">//storing &class="macro">#x27;None&class="macro">#x27; as the previous value } class=class="str">"cmt">//-- storing the event impact from(ENUM_CALENDAR_EVENT_IMPACT) to a class="type">class="kw">string NewsTime[Size].EventImpact = EnumToString(values[x].impact_type); class=class="str">"cmt">//-- storing the event frequency from(ENUM_CALENDAR_EVENT_FREQUENCY) to a class="type">class="kw">string NewsTime[Size].EventFrequency = EnumToString(event.frequency); Size++;class=class="str">"cmt">//incrementing the Calendar array NewsTime
「用 sqlite_master 给财经库做结构对账」
| 在 MT5 里做新闻/财经日历的本地库同步，第一步不是直接写表，而是先读 sqlite_master 摸清库里已经有什么。下面这段逻辑打开公共目录下的 NEWS_DATABASE_FILE，用 DATABASE_OPEN_READWRITE | DATABASE_OPEN_CREATE | DATABASE_OPEN_COMMON 三标志，文件不存在且句柄无效时直接返回 true，意味着首次运行允许后续建表。 |
|---|
打开成功后走 DatabasePrepare 发一句 select * from sqlite_master where type<>'index';，把非索引对象全部捞出来。循环里用 DatabaseReadBind 往 SQLiteMaster 结构读，每读一行就 ArrayResize(DBContents,i+1,i+2) 动态扩数组——这种按 2 增长的上限扩容，在记录数波动大时比每次 +1 少几次重分配。
对账环节把 CalendarContents 里每句的 IF NOT EXISTS 用 StringReplace 抹掉再比 sql 文本，因为建表语句本身带不带这个子句不影响表结构是否一致。DBContents[i].sql 还会检查末尾有没有分号，没有就补一个，避免拼接执行时语法断在半句。
外汇与贵金属事件驱动策略依赖这类本地库，但财经数据延迟或结构变更都可能让比对失效，实战中需以经纪商实际推送为准，相关交易本身属高风险。
class="type">bool CNews::UpdateRecords() { class=class="str">"cmt">//initialize variable to true class="type">bool perform_update=true; class=class="str">"cmt">//--- open/create class=class="str">"cmt">//-- try to open database Calendar class="type">int db=DatabaseOpen(NEWS_DATABASE_FILE, DATABASE_OPEN_READWRITE | DATABASE_OPEN_CREATE| DATABASE_OPEN_COMMON); if(db==INVALID_HANDLE)class=class="str">"cmt">//Checks if the database was able to be opened { class=class="str">"cmt">//if opening the database failed if(!FileIsExist(NEWS_DATABASE_FILE,FILE_COMMON))class=class="str">"cmt">//Checks if the database Calendar exists in the common folder { class="kw">return perform_update;class=class="str">"cmt">//Returns true when the database was failed to be opened and the file doesn&class="macro">#x27;t exist in the common folder } } class="type">int MasterRequest = DatabasePrepare(db,"select * from sqlite_master where type<>&class="macro">#x27;index&class="macro">#x27;;"); if(MasterRequest==INVALID_HANDLE) { Print("DB: ",NEWS_DATABASE_FILE, " request failed with code ", GetLastError()); } else { SQLiteMaster ReadContents; class=class="str">"cmt">//Assigning values from the sql query into DBContents array for(class="type">int i=class="num">0; DatabaseReadBind(MasterRequest,ReadContents); i++) { ArrayResize(DBContents,i+class="num">1,i+class="num">2); DBContents[i].type = ReadContents.type; DBContents[i].name = ReadContents.name; DBContents[i].tbl_name = ReadContents.tbl_name; DBContents[i].rootpage = ReadContents.rootpage; class=class="str">"cmt">/*Check if the end of the sql class="type">class="kw">string has a character &class="macro">#x27;;&class="macro">#x27; if not add this character to the class="type">class="kw">string*/ DBContents[i].sql = (StringFind(ReadContents.sql,";",StringLen(ReadContents.sql)-class="num">1)== (StringLen(ReadContents.sql)-class="num">1))?ReadContents.sql:ReadContents.sql+";";; } class="type">uint contents_exists = class="num">0; for(class="type">uint i=class="num">0;i<DBContents.Size();i++) { class="type">bool isCalendarContents = false; for(class="type">uint x=class="num">0;x<CalendarContents.Size();x++) { class=class="str">"cmt">/*Store Sql query from CalendarContents without class="type">class="kw">string &class="macro">#x27; IF NOT EXISTS&class="macro">#x27;*/ class="type">class="kw">string CalendarSql=CalendarContents[x].sql; StringReplace(CalendarSql," IF NOT EXISTS",""); class=class="str">"cmt">//-- Check if the Db object is in our list if(DBContents[i].name==CalendarContents[x].name&& (DBContents[i].sql==CalendarSql||
◍ 用 SQL 比对给财经日历库做增量更新判定
这段逻辑核心是解决本地财经日历数据库和远端日历内容的一致性：先遍历已存在的 DBContents，拿 sql、type、tbl_name 三个字段和 CalendarContents 逐条比对，命中就累加 contents_exists 并置 isCalendarContents 为 true。
没匹配上的本地表直接被 Drop 掉，终端会打印类似 DBContent: xxx is not needed! 的提示，顺带输出 Attempting To Clean Database...，相当于自动清垃圾表。
一致性门槛很硬：只有当 contents_exists 等于 CalendarContents.Size() 时，才认为库结构齐备；差一个就返回 perform_update 让上层去拉新数据。
最后一步是查 Record 表里的日期——用 SELECT Date FROM Record where Date=Date(REPLACE('2024.05.21','.','-')) 这种写法把交易服务器时间转成 ISO 格式去做相等判断。若当天已存在记录，perform_update 置 false 跳过更新；读不到或请求句柄 INVALID_HANDLE 就倾向触发更新，错误码靠 GetLastError() 查。外汇与贵金属事件驱动波动大，这类自动更新判定出错可能让你漏掉关键风险事件，上 MT5 跑前先手测一次日期格式化。
DBContents[i].sql==CalendarContents[x].sql)&&
CalendarContents[x].type==DBContents[i].type&&
CalendarContents[x].tbl_name==DBContents[i].tbl_name)
{
contents_exists++;
isCalendarContents = true;
}
}
if(!isCalendarContents)
{
class=class="str">"cmt">//-- Print DBcontent&class="macro">#x27;s name if it does not match with CalendarContents
PrintFormat("DBContent: %s is not needed!",DBContents[i].name);
class=class="str">"cmt">//-- We will drop the table if it is not neccessary
DatabaseExecute(db,StringFormat(DropRequest,DBContents[i].type,DBContents[i].name));
Print("Attempting To Clean Database...");
}
}
class=class="str">"cmt">/*If not all the CalendarContents exist in the Calendar Database before an update */
if(contents_exists!=CalendarContents.Size())
{
class="kw">return perform_update;
}
}
if(!DatabaseTableExists(db,CalendarStruct(Record_Table).name))class=class="str">"cmt">//If the database table &class="macro">#x27;Record&class="macro">#x27; doesn&class="macro">#x27;t exist
{
DatabaseClose(db);
class="kw">return perform_update;
}
class=class="str">"cmt">//-- Sql query to determine the lastest or maximum date recorded
class=class="str">"cmt">/* If the last recorded date data in the &class="macro">#x27;Record&class="macro">#x27; table is not equal to the current day, perform an update! */
class="type">class="kw">string request_text=StringFormat("SELECT Date FROM %s where Date=Date(REPLACE(&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;.&class="macro">#x27;,&class="macro">#x27;-&class="macro">#x27;))",
CalendarStruct(Record_Table).name,TimeToString(TimeTradeServer()));
class="type">int request=DatabasePrepare(db,request_text);class=class="str">"cmt">//Creates a handle of a request, which can then be executed using DatabaseRead()
if(request==INVALID_HANDLE)class=class="str">"cmt">//Checks if the request failed to be completed
{
Print("DB: ",NEWS_DATABASE_FILE, " request failed with code ", GetLastError());
DatabaseClose(db);
class="kw">return perform_update;
}
if(DatabaseRead(request))class=class="str">"cmt">//Will be true if there are results from the sql query/request
{
DatabaseFinalize(request);class=class="str">"cmt">//Removes a request created in DatabasePrepare()
DatabaseClose(db);class=class="str">"cmt">//Closes the database
perform_update=false;
class="kw">return perform_update;
}
else
{
DatabaseFinalize(request);class=class="str">"cmt">//Removes a request created in DatabasePrepare()
DatabaseClose(db);class=class="str">"cmt">//Closes the database
class="kw">return perform_update;
}
}经纪商财经日历库的本地建表逻辑
在 MT5 里做新闻驱动的贵金属或外汇策略，第一步往往是把经纪商的经济日历落到本地 SQLite，避免每次都去拉在线接口。下面这段 CNews::CreateEconomicDatabase 就是干这件事的入口：先判断库文件是否存在，存在就走 UpdateRecords 增量更新，不存在才新建。 它用 DatabaseOpen 以 DATABASE_OPEN_COMMON 标志在公共目录开库，这意味着多终端账号能共享同一份日历数据；同时另开一个 NEWS_TEXT_FILE 文本锁文件，建表成功前若任意一步失败（比如 DatabaseOpen 返回 INVALID_HANDLE 或 FileOpen 写不了），就删锁文件并回滚，防止半截库被误用。 建表前用 DatabaseTransactionBegin 包住，两张表 MQL5Calendar 与 TimeSchedule 通过 CreateCalendarTable / CreateTimeTable 尝试创建；若 tableExists 为真只打印 Updating 提示，否则视为首建。整套流程跑完前终端会卡在 Please wait，实盘加载 EA 时要注意这个阻塞窗口。外汇与贵金属受数据行情跳空影响大，本地库只是降低延迟，不消除风险。
class="type">void CNews::CreateEconomicDatabase() { if(FileIsExist(NEWS_DATABASE_FILE,FILE_COMMON))class=class="str">"cmt">//Check if the database exists { if(!UpdateRecords())class=class="str">"cmt">//Check if the database is up to date { class="kw">return;class=class="str">"cmt">//will terminate execution of the rest of the code below } } if(FileIsExist(NEWS_TEXT_FILE,FILE_COMMON))class=class="str">"cmt">//Check if the database is open { class="kw">return;class=class="str">"cmt">//will terminate execution of the rest of the code below } Calendar Evalues[];class=class="str">"cmt">//Creating a Calendar array variable class="type">bool failed=false,tableExists=false; class="type">int file=INVALID_HANDLE; class=class="str">"cmt">//--- open/create the database &class="macro">#x27;Calendar&class="macro">#x27; class=class="str">"cmt">//-- will try to open/create in the common folder class="type">int db=DatabaseOpen(NEWS_DATABASE_FILE, DATABASE_OPEN_READWRITE | DATABASE_OPEN_CREATE| DATABASE_OPEN_COMMON); if(db==INVALID_HANDLE)class=class="str">"cmt">//Checks if the database &class="macro">#x27;Calendar&class="macro">#x27; failed to open/create { Print("DB: ",NEWS_DATABASE_FILE, " open failed with code ", GetLastError()); class="kw">return;class=class="str">"cmt">//will terminate execution of the rest of the code below } else { class=class="str">"cmt">//-- try to create a text file &class="macro">#x27;NewsDatabaseOpen&class="macro">#x27; in common folder file=FileOpen(NEWS_TEXT_FILE,FILE_WRITE|FILE_ANSI|FILE_TXT|FILE_COMMON); if(file==INVALID_HANDLE) { DatabaseClose(db);class=class="str">"cmt">//Closes the database &class="macro">#x27;Calendar&class="macro">#x27; if the News text file failed to be created class="kw">return;class=class="str">"cmt">//will terminate execution of the rest of the code below } } DatabaseTransactionBegin(db);class=class="str">"cmt">//Starts transaction execution Print("Please wait..."); class=class="str">"cmt">//-- attempt to create the MQL5Calendar and TimeSchedule tables if(!CreateCalendarTable(db,tableExists)||!CreateTimeTable(db,tableExists)) { FileClose(file);class=class="str">"cmt">//Closing the file &class="macro">#x27;NewsDatabaseOpen.txt&class="macro">#x27; FileDelete(NEWS_TEXT_FILE,FILE_COMMON);class=class="str">"cmt">//Deleting the file &class="macro">#x27;NewsDatabaseOpen.txt&class="macro">#x27; class="kw">return;class=class="str">"cmt">//will terminate execution of the rest of the code below } EconomicDetails(Evalues);class=class="str">"cmt">//Retrieving the data from the Economic Calendar if(tableExists)class=class="str">"cmt">//Checks if there is an existing table within the Calendar Database { class=class="str">"cmt">//if there is an existing table we will notify the user that we are updating the table. PrintFormat("Updating %s",NEWS_DATABASE_FILE); } else {
「建表失败时的回滚与成功后的视图落地」
往日历库写经济事件数据，最怕中途插入失败却留下半截脏数据。这段逻辑用 failed 标志位判断 InsertIntoTables 的返回，一旦为真立刻 DatabaseTransactionRollback(db) 解锁回滚，同时 FileClose 关掉 NEWS_TEXT_FILE 文本文件并 FileDelete 删掉它，再用 ArrayRemove(Evalues,0,WHOLE_ARRAY) 清空内存数组，整个写入尝试算作废。 若 failed 为 false，说明 MQL5Calendar 与 TimeSchedule 两张表都写进去了，接着跑 CreateCalendarViews 建视图、CreateRecordTable 记当前时间、CreateAutoDST 存券商 DST 排程；文本文件和数组同样清理掉。tableExists 若为 true 终端打印「库名 Updated」，否则打印「库名 Created」——交易者开 MT5 看日志就能确认本次是更新还是新建。 收尾不论哪条分支，都走 DatabaseTransactionCommit(db) 落盘再 DatabaseClose(db) 关库。CreateCalendarTable 里先用 DatabaseTableExists 查 MQL5Calendar 是否存在，存在则 tableExists=true，并继续查 TimeSchedule；若 TimeSchedule 也在，注释写明会 drop 掉旧表重建，避免结构不一致。外汇与贵金属事件驱动策略依赖这套时序数据，但行情跳空与滑点属高风险，回测结论仅代表历史概率。
if(!InsertIntoTables(db,Evalues)) { failed=true; } if(failed) { DatabaseTransactionRollback(db); PrintFormat("%s: DatabaseExecute() failed with code %d", __FUNCTION__, GetLastError()); FileClose(file); FileDelete(NEWS_TEXT_FILE,FILE_COMMON); ArrayRemove(Evalues,class="num">0,WHOLE_ARRAY); } else { CreateCalendarViews(db); CreateRecordTable(db); CreateAutoDST(db); FileClose(file); FileDelete(NEWS_TEXT_FILE,FILE_COMMON); ArrayRemove(Evalues,class="num">0,WHOLE_ARRAY); if(tableExists) PrintFormat("%s Updated",NEWS_DATABASE_FILE); else PrintFormat("%s Created",NEWS_DATABASE_FILE); } DatabaseTransactionCommit(db); DatabaseClose(db); class="type">bool CNews::CreateCalendarTable(class="type">int db,class="type">bool &tableExists) { if(DatabaseTableExists(db,CalendarStruct(MQL5Calendar_Table).name)) { tableExists=true; if(DatabaseTableExists(db,CalendarStruct(TimeSchedule_Table).name)) { class=class="str">"cmt">//-- We will drop the table if the table already exists
◍ 建表前的删表兜底与失败退出
在 MT5 用 DatabaseExecute 操作 SQLite 时，若目标表已存在又不想留脏数据，标准做法是先发一条 Drop Table 再建表。下面两段对 TimeSchedule 与 MQL5Calendar 都做了存在即删的处理，删失败就直接 DatabaseClose 并 return false，避免后续逻辑在残缺库结构上继续跑。 if(!DatabaseExecute(db,StringFormat("Drop Table %s",CalendarStruct(TimeSchedule_Table).name))) { PrintFormat("Failed to drop table %s with code %d",CalendarStruct(TimeSchedule_Table).name,GetLastError()); DatabaseClose(db); return false; } 上面这段里 StringFormat 把表名拼进 SQL，GetLastError 返回的是本次执行的错误码（如 5120 类数据库错误），PrintFormat 把它打进日志方便定位。 删完表后，用 DatabaseTableExists 判断 MQL5Calendar 是否真的不在，不在才用 CalendarStruct(...).sql 建表；建表失败同样关库返 false。CreateTimeTable 函数对 TimeSchedule 也套同一套逻辑，tableExists 引用参数在这里没被改写，调用方若想知表是否已存在得自己再查一次。 开 MT5 把这段塞进 EA 的数据库初始化函数，故意先把表手动建好再跑，能在日志里看到 Drop 成功；把 db 句柄改成无效值就能验证返 false 分支。外汇与贵金属策略跑这类库操作仍属高风险，任何库异常都可能让定时器漏触发。
if(!DatabaseExecute(db,StringFormat("Drop Table %s",CalendarStruct(TimeSchedule_Table).name))) { class=class="str">"cmt">//If the table failed to be dropped/deleted PrintFormat("Failed to drop table %s with code %d",CalendarStruct(TimeSchedule_Table).name,GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close the database class="kw">return false;class=class="str">"cmt">//will terminate execution of the rest of the code below and class="kw">return false, when the table cannot be dropped } class=class="str">"cmt">//--We will drop the table if the table already exists if(!DatabaseExecute(db,StringFormat("Drop Table %s",CalendarStruct(MQL5Calendar_Table).name))) { class=class="str">"cmt">//If the table failed to be dropped/deleted PrintFormat("Failed to drop table %s with code %d",CalendarStruct(MQL5Calendar_Table).name,GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close the database class="kw">return false;class=class="str">"cmt">//will terminate execution of the rest of the code below and class="kw">return false, when the table cannot be dropped } class=class="str">"cmt">//-- If the database table &class="macro">#x27;MQL5Calendar&class="macro">#x27; doesn&class="macro">#x27;t exist if(!DatabaseTableExists(db,CalendarStruct(MQL5Calendar_Table).name)) { class=class="str">"cmt">//--- create the table &class="macro">#x27;MQL5Calendar&class="macro">#x27; if(!DatabaseExecute(db,CalendarStruct(MQL5Calendar_Table).sql))class=class="str">"cmt">//Checks if the table was successfully created { Print("DB: create the Calendar table failed with code ", GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close the database class="kw">return false;class=class="str">"cmt">//Function returns false if creating the table failed } } class="kw">return true;class=class="str">"cmt">//Function returns true if creating the table was successful } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Function for creating a table in a database | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">bool CNews::CreateTimeTable(class="type">int db,class="type">bool &tableExists) { class=class="str">"cmt">//-- If the database table &class="macro">#x27;TimeSchedule&class="macro">#x27; doesn&class="macro">#x27;t exist if(!DatabaseTableExists(db,CalendarStruct(TimeSchedule_Table).name)) { class=class="str">"cmt">//--- create the table &class="macro">#x27;TimeSchedule&class="macro">#x27; if(!DatabaseExecute(db,CalendarStruct(TimeSchedule_Table).sql))class=class="str">"cmt">//Checks if the table was successfully created { Print("DB: create the Calendar table failed with code ", GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close the database class="kw">return false;class=class="str">"cmt">//Function returns false if creating the table failed } } class="kw">return true;class=class="str">"cmt">//Function returns true if creating the table was successful }
把财经日历落库时顺手调好英美夏令时
做 MT5 财经数据本地化，最容易被忽略的是时区漂移。上面这段把经济事件写进数据库时，对 EventDate 做了 UK DST 与 US DST 两次 adjustDaylightSavings 校正，否则同一事件在夏时制切换前后可能差 1 小时，回测时容易错配 K 线。 CreateCalendarViews 只循环 i=1 到 4，对应四种日历视图组件，任一个 DatabaseExecute 返回失败就 Print 出 GetLastError 代码，方便你直接在专家日志里抓建视图失败的环节。 InsertIntoTables 用 StringFormat 把 Calendar 数组逐条拼成 INSERT，字段从 EventId 到 EventFrequency 共 14 个，循环到 Evalues.Size() 为止。若你接的是 XAUUSD 或 EURUSD 这类受英美数据双重影响的品种，落库前不校正夏令时，事件驱动策略的触发点可能整体偏移。 外汇与贵金属波动受数据面冲击大、杠杆风险高，写库逻辑建议先在策略测试器用小规模 Evalues 跑通再上实盘。
class="type">void CNews::CreateCalendarViews(class="type">int db) { for(class="type">uint i=class="num">1;i<=class="num">4;i++) { if(!DatabaseExecute(db,CalendarStruct((CalendarComponents)i).sql))class=class="str">"cmt">//Checks if the view was successfully created { Print("DB: create the Calendar view failed with code ", GetLastError()); } } } class="type">bool CNews::InsertIntoTables(class="type">int db,Calendar &Evalues[]) { for(class="type">uint i=class="num">0; i<Evalues.Size(); i++)class=class="str">"cmt">//Looping through all the Economic Events { class="type">class="kw">string request_insert_into_calendar = StringFormat(CalendarStruct(MQL5Calendar_Table).insert, i, Evalues[i].EventId, Evalues[i].CountryName, Evalues[i].EventName, Evalues[i].EventType, Evalues[i].EventImportance, Evalues[i].EventCurrency, Evalues[i].EventCode, Evalues[i].EventSector, Evalues[i].EventForecast, Evalues[i].EventPreval, Evalues[i].EventImpact, Evalues[i].EventFrequency);class=class="str">"cmt">//Inserting all the columns for each event record if(DatabaseExecute(db,request_insert_into_calendar))class=class="str">"cmt">//Check if insert query into calendar was successful { class="type">class="kw">string request_insert_into_time = StringFormat(CalendarStruct(TimeSchedule_Table).insert, i, class=class="str">"cmt">//-- Economic EventDate adjusted for UK DST(Daylight Savings Time) Savings_UK.adjustDaylightSavings(StringToTime(Evalues[i].EventDate)), class=class="str">"cmt">//-- Economic EventDate adjusted for US DST(Daylight Savings Time)
「夏令时校正与更新记录表的落地写法」
把财经事件写进本地 SQLite 时，美澳夏令时偏移必须分别用 Savings_US / Savings_AU 两个实例的 adjustDaylightSavings 处理，否则同一 EventDate 在悉尼与纽约会差出半小时到一小时，回测时事件触发点会整体错位。 插入失败不要静默跳过：DatabaseExecute 返回 false 时立刻 Print(GetLastError()) 并把拼好的 SQL 原样打印，能直接看到是字段默认值缺失还是引号转义崩了。循环里任一条插不进去就 return false，避免半张表有数据、半张表空着还报成功。 CreateRecordTable 负责建一张记录“库最后更新时间”的表。先用 DatabaseTableExists 判存在，不存在才跑建表 SQL；建完无论走哪条分支都执行一次 Record_Trigger 的 SQL，保证时间戳自写触发器一直在。 别把正态当圣经：MT5 内置 DatabaseExecute 对并发写不保序，若小布盯盘脚本和多周期 EA 同时写 Calendar 库，可能丢更新记录，实盘前用 Print 出来的最后写入时间对一遍。
Savings_US.adjustDaylightSavings(StringToTime(Evalues[i].EventDate)), class=class="str">"cmt">//-- Economic EventDate adjusted for AU DST(Daylight Savings Time) Savings_AU.adjustDaylightSavings(StringToTime(Evalues[i].EventDate)), Evalues[i].EventDateclass=class="str">"cmt">//normal Economic EventDate );class=class="str">"cmt">//Inserting all the columns for each event record if(!DatabaseExecute(db,request_insert_into_time)) { Print(GetLastError()); class=class="str">"cmt">//-- Will print the sql query to check for any errors or possible defaults in the query/request Print(request_insert_into_time); class="kw">return false;class=class="str">"cmt">//Will end the loop and class="kw">return false, as values failed to be inserted into the table } } else { Print(GetLastError()); class=class="str">"cmt">//-- Will print the sql query to check for any errors or possible defaults in the query/request Print(request_insert_into_calendar); class="kw">return false;class=class="str">"cmt">//Will end the loop and class="kw">return false, as values failed to be inserted into the table } } class="kw">return true;class=class="str">"cmt">//Will class="kw">return true, all values were inserted into the table successfully } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Creates a table to store the record of when last the Calendar | class=class="str">"cmt">//|database was updated/created | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">void CNews::CreateRecordTable(class="type">int db) { class="type">bool failed=false; if(!DatabaseTableExists(db,CalendarStruct(Record_Table).name))class=class="str">"cmt">//Checks if the table &class="macro">#x27;Record&class="macro">#x27; exists in the databse &class="macro">#x27;Calendar&class="macro">#x27; { class=class="str">"cmt">//--- create the table if(!DatabaseExecute(db,CalendarStruct(Record_Table).sql))class=class="str">"cmt">//Will attempt to create the table &class="macro">#x27;Record&class="macro">#x27; { Print("DB: create the Records table failed with code ", GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close the database class="kw">return;class=class="str">"cmt">//Exits the function if creating the table failed } elseclass=class="str">"cmt">//If Table was created Successfully then Create Trigger { DatabaseExecute(db,CalendarStruct(Record_Trigger).sql); } } else { DatabaseExecute(db,CalendarStruct(Record_Trigger).sql); } class=class="str">"cmt">//Sql query/request to insert the current time into the &class="macro">#x27;Date&class="macro">#x27; column in the table &class="macro">#x27;Record&class="macro">#x27;
◍ 把经纪商夏令时写进本地库的事务兜底
这段逻辑干的事很实在：先靠 AutoDetectDST() 跑完时区推断流程，失败就直接 return 不往下走，避免脏数据落库。 若 Calendar 库里还没有 AutoDST 表，就先用 CalendarStruct(AutoDST_Table).sql 建表；建表成功才顺手建触发器，建表失败则关库退出。表已存在时则跳过建表、只补触发器。 插入动作靠 StringFormat 把 DSTType 枚举拼成 SQL：DatabaseExecute 若返回 false，先把 GetLastError 和完整 SQL 打到日志，再把 failed 置 true。 failed 为真时调用 DatabaseTransactionRollback 回滚全部事务并释放锁——外汇与贵金属行情受经纪商 DST 切换影响，时区错乱可能导致回测与实盘时间基准偏移，这类落库操作需在 MT5 里用小样本先验证回滚是否生效。
class="type">class="kw">string request_text=StringFormat(CalendarStruct(Record_Table).insert,TimeToString(TimeTradeServer())); if(!DatabaseExecute(db, request_text))class=class="str">"cmt">//Will attempt to run this sql request/query { Print(GetLastError()); PrintFormat(CalendarStruct(Record_Table).insert,TimeToString(TimeTradeServer())); failed=true;class=class="str">"cmt">//assign true if the request failed } if(failed) { class=class="str">"cmt">//--- roll back all transactions and unlock the database DatabaseTransactionRollback(db); PrintFormat("%s: DatabaseExecute() failed with code %d", __FUNCTION__, GetLastError()); } } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Function for creating and inserting Recommend DST for the Broker | class=class="str">"cmt">//|into a table | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">void CNews::CreateAutoDST(class="type">int db) { class="type">bool failed=false;class=class="str">"cmt">//boolean variable if(!AutoDetectDST(DSTType))class=class="str">"cmt">//Check if AutoDetectDST went through all the right procedures { class="kw">return;class=class="str">"cmt">//will terminate execution of the rest of the code below } if(!DatabaseTableExists(db,CalendarStruct(AutoDST_Table).name))class=class="str">"cmt">//Checks if the table &class="macro">#x27;AutoDST&class="macro">#x27; exists in the databse &class="macro">#x27;Calendar&class="macro">#x27; { class=class="str">"cmt">//--- create the table AutoDST if(!DatabaseExecute(db,CalendarStruct(AutoDST_Table).sql))class=class="str">"cmt">//Will attempt to create the table &class="macro">#x27;AutoDST&class="macro">#x27; { Print("DB: create the AutoDST table failed with code ", GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close the database class="kw">return;class=class="str">"cmt">//Exits the function if creating the table failed } elseclass=class="str">"cmt">//If Table was created Successfully then Create Trigger { DatabaseExecute(db,CalendarStruct(AutoDST_Trigger).sql); } } else { class=class="str">"cmt">//Create trigger if AutoDST table exists DatabaseExecute(db,CalendarStruct(AutoDST_Trigger).sql); } class=class="str">"cmt">//Sql query/request to insert the recommend DST for the Broker using the DSTType variable to determine which class="type">class="kw">string data to insert class="type">class="kw">string request_text=StringFormat(CalendarStruct(AutoDST_Table).insert,EnumToString(DSTType)); if(!DatabaseExecute(db, request_text))class=class="str">"cmt">//Will attempt to run this sql request/query { Print(GetLastError()); PrintFormat(CalendarStruct(AutoDST_Table).insert,EnumToString(DSTType));class=class="str">"cmt">//Will print the sql query if failed failed=true;class=class="str">"cmt">//assign true if the request failed } if(failed) { class=class="str">"cmt">//--- roll back all transactions and unlock the database DatabaseTransactionRollback(db);
从财经库里抠出最近更新时间
做新闻日历类 EA 时，先确认本地缓存库里最新一条记录的时间，才能决定要不要拉新数据。下面这段从公共目录只读打开 Calendar 库，若文件不存在直接回退到 1970.01.01 00:00:00，避免后面比较逻辑炸掉。 核心 SQL 只有一句：SELECT REPLACE(Date,'-','.') FROM 'Record'，把日期里的横杠换成点号以适配 MQL5 的 datetime 字符串解析。读出来的文本赋值给 latest_record，函数末尾用 (datetime)latest_record 强转返回。 如果 DatabasePrepare 或 DatabaseRead 任一环节拿到 INVALID_HANDLE，就打印 GetLastError() 错误码并顺手 DatabaseFinalize + DatabaseClose，否则句柄泄漏会在跑几十次后让 MT5 终端变卡。外汇与贵金属消息面波动剧烈，这类缓存时间误差可能导致错过低相关性行情，实盘前务必在策略测试器里手动改几条 Record 验证边界。
class="type">class="kw">datetime CNews::GetLatestNewsDate() { class=class="str">"cmt">//--- open the database &class="macro">#x27;Calendar&class="macro">#x27; in the common folder class="type">int db=DatabaseOpen(NEWS_DATABASE_FILE, DATABASE_OPEN_READONLY|DATABASE_OPEN_COMMON); if(db==INVALID_HANDLE)class=class="str">"cmt">//Checks if &class="macro">#x27;Calendar&class="macro">#x27; failed to be opened { if(!FileIsExist(NEWS_DATABASE_FILE,FILE_COMMON))class=class="str">"cmt">//Checks if &class="macro">#x27;Calendar&class="macro">#x27; database exists { Print("Could not find Database!"); class="kw">return class="num">0;class=class="str">"cmt">//Will class="kw">return the earliest date which is class="num">1970.01.class="num">01 class="num">00:class="num">00:class="num">00 } } class="type">class="kw">string latest_record="class="num">1970.01.class="num">01";class=class="str">"cmt">//class="type">class="kw">string variable with the first/earliest possible date in MQL5 class=class="str">"cmt">//Sql query to determine the lastest or maximum recorded time from which the database was updated. class="type">class="kw">string request_text="SELECT REPLACE(Date,&class="macro">#x27;-&class="macro">#x27;,&class="macro">#x27;.&class="macro">#x27;) FROM &class="macro">#x27;Record&class="macro">#x27;"; class="type">int request=DatabasePrepare(db,request_text); if(request==INVALID_HANDLE) { Print("DB: ",NEWS_DATABASE_FILE, " request failed with code ", GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close Database class="kw">return class="num">0; } if(DatabaseRead(request))class=class="str">"cmt">//Will read the one record in the &class="macro">#x27;Record&class="macro">#x27; table { class=class="str">"cmt">//-- Will assign the first column(column class="num">0) value to the variable &class="macro">#x27;latest_record&class="macro">#x27; if(!DatabaseColumnText(request,class="num">0,latest_record)) { Print("DatabaseRead() failed with code ", GetLastError()); DatabaseFinalize(request);class=class="str">"cmt">//Finalize request DatabaseClose(db);class=class="str">"cmt">//Closes the database &class="macro">#x27;Calendar&class="macro">#x27; class="kw">return D&class="macro">#x27;class="num">1970.01.class="num">01&class="macro">#x27;;class=class="str">"cmt">//Will end the for loop and will class="kw">return the earliest date which is class="num">1970.01.class="num">01 class="num">00:class="num">00:class="num">00 } } DatabaseFinalize(request); DatabaseClose(db);class=class="str">"cmt">//Closes the database &class="macro">#x27;Calendar&class="macro">#x27; class="kw">return (class="type">class="kw">datetime)latest_record;class=class="str">"cmt">//Returns the class="type">class="kw">string latest_record converted to class="type">class="kw">datetime }
「抓去年美国NFP日期并定位经纪商EURUSD」
做跨年非农回测时，先得把上一年度的日历边界算准。下面这段代码用 Time.ReturnYear 配合偏移接口拿到去年年份，再拼出 01.01 00:00:00 到 12.31 23:59:59 的起止 datetime，喂给 CalendarValueHistory 拉取 CountryCode='US' 的全部日历值。 在返回数组里只筛 event_id==840030016 的记录，这是 NFP（非农就业）的事件 ID。每命中一条就把其时间转成字符串压进 eventtime 数组，这样你就拿到了去年全部 NFP 公布时刻，精度到秒。 光有日期不够，还得绑定具体品种。代码随后遍历 Market Watch 内全部 Symbol，用基础货币 EUR、计价货币 USD 且非自定义品种的条件去抓真实的 EURUSD 报价名——不同经纪商后缀可能带 .pro 或 .m，直接写死 "EURUSD" 在部分终端会取不到句柄。 若观察列表里没找到，就再扫一遍 SymbolsTotal(false) 的全市场品种补兜底。外汇和贵金属受非农跳空影响明显，杠杆品种风险高，这段代码只解决“取数+绑品种”，实盘或回测请自行校验点差与滑点。
class="type">int lastyear = Time.ReturnYear(Time.TimeMinusOffset(iTime(Symbol(),PERIOD_CURRENT,class="num">0),Time.Years())); class=class="str">"cmt">//-- Will store the start date for the previous year class="type">class="kw">datetime lastyearstart = StringToTime(StringFormat("%s.class="num">01.01 class="num">00:class="num">00:class="num">00",(class="type">class="kw">string)lastyear)); class=class="str">"cmt">//-- Will store the end date for the previous year class="type">class="kw">datetime lastyearend = StringToTime(StringFormat("%s.class="num">12.31 class="num">23:class="num">59:class="num">59",(class="type">class="kw">string)lastyear)); class=class="str">"cmt">//-- Getting last year&class="macro">#x27;s calendar values for CountryCode = &class="macro">#x27;US&class="macro">#x27; if(CalendarValueHistory(values,lastyearstart,lastyearend,"US")) { for(class="type">int x=class="num">0; x<(class="type">int)ArraySize(values); x++) { if(values[x].event_id==class="num">840030016)class=class="str">"cmt">//Get only NFP Event Dates { ArrayResize(eventtime,eventtime.Size()+class="num">1,eventtime.Size()+class="num">2);class=class="str">"cmt">//Increasing the size of eventtime array by class="num">1 eventtime[eventtime.Size()-class="num">1] = TimeToString(values[x].time);class=class="str">"cmt">//Storing the dates in an array of type class="type">class="kw">string } } } class=class="str">"cmt">//-- class="type">class="kw">datetime variables to store the broker&class="macro">#x27;s timezone shift(change) class="type">class="kw">datetime ShiftStart=D&class="macro">#x27;class="num">1970.01.class="num">01 class="num">00:class="num">00:class="num">00&class="macro">#x27;,ShiftEnd=D&class="macro">#x27;class="num">1970.01.class="num">01 class="num">00:class="num">00:class="num">00&class="macro">#x27;; class="type">class="kw">string EURUSD="";class=class="str">"cmt">//String variables declarations for working with EURUSD class="type">bool EurusdIsFound=false;class=class="str">"cmt">//Boolean variables declarations for working with EURUSD for(class="type">int i=class="num">0;i<SymbolsTotal(true);i++)class=class="str">"cmt">//Will loop through all the Symbols inside the Market Watch { class="type">class="kw">string SymName = SymbolName(i,true);class=class="str">"cmt">//Assign the Symbol Name of index &class="macro">#x27;i&class="macro">#x27; from the list of Symbols inside the Market Watch class=class="str">"cmt">//-- Check if the Symbol outside the Market Watch has a SYMBOL_CURRENCY_BASE of EUR class=class="str">"cmt">//-- and a SYMBOL_CURRENCY_PROFIT of USD, and this Symbol is not a Custom Symbol(Is not from the broker) if(((CurrencyBase(SymName)=="EUR"&&CurrencyProfit(SymName)=="USD")|| (StringFind(SymName,"EUR")>-class="num">1&&CurrencyProfit(SymName)=="USD"))&&!Custom(SymName)) { EURUSD = SymName;class=class="str">"cmt">//Assigning the name of the EURUSD Symbol found inside the Market Watch EurusdIsFound = true;class=class="str">"cmt">//EURUSD Symbol was found in the Trading Terminal for your Broker class="kw">break;class=class="str">"cmt">//Will end the for loop } } if(!EurusdIsFound)class=class="str">"cmt">//Check if EURUSD Symbol was already Found in the Market Watch { for(class="type">int i=class="num">0; i<SymbolsTotal(false); i++)class=class="str">"cmt">//Will loop through all the available Symbols outside the Market Watch
◍ 用 EURUSD 异动反推券商时区切换
这段逻辑先在市场报价之外的全部品种里捞一个 EURUSD：基础货币是 EUR、盈利货币是 USD，且不是自定义品种（非券商外来的假符号）；若名称里带 EUR 且盈利货币为 USD 也认。找到后跳出循环，找不到就打印三条提示并把 dstType 置为 DST_NONE 直接返回 false，意味着该券商无夏令时调度可供识别。 接下来用结构体 DST 存每次历史事件日的判定结果（是否比前后一根都大的 K 线），遍历 eventtime 数组。若当前结果与上一次不同，就认为时间发生了偏移，timeIsShifted 置真。 偏移起点 ShiftStart 初次命中时记录当前事件日，ShiftEnd 则记为上一次事件日。这样你开 MT5 把这段塞进 EA 初始化，能在券商不改公开时区说明的情况下，凭 EURUSD 成交量异常日反推出夏令时切换边界，外汇与贵金属杠杆品种受此影响报价跳空，属高风险验证项。
{
class="type">class="kw">string SymName = SymbolName(i,false);class=class="str">"cmt">//Assign the Symbol Name of index &class="macro">#x27;i&class="macro">#x27; from the list of Symbols outside the Market Watch
class=class="str">"cmt">//-- Check if the Symbol outside the Market Watch has a SYMBOL_CURRENCY_BASE of EUR
class=class="str">"cmt">//-- and a SYMBOL_CURRENCY_PROFIT of USD, and this Symbol is not a Custom Symbol(Is not from the broker)
if(((CurrencyBase(SymName)=="EUR"&&CurrencyProfit(SymName)=="USD")||
(StringFind(SymName,"EUR")>-class="num">1&&CurrencyProfit(SymName)=="USD"))&&!Custom(SymName))
{
EURUSD = SymName;class=class="str">"cmt">//Assigning the name of the EURUSD Symbol found outside the Market Watch
EurusdIsFound = true;class=class="str">"cmt">//EURUSD Symbol was found in the Trading Terminal for your Broker
class="kw">break;class=class="str">"cmt">//Will end the for loop
}
}
}
if(!EurusdIsFound)class=class="str">"cmt">//Check if EURUSD Symbol was Found in the Trading Terminal for your Broker
{
Print("Cannot Find EURUSD!");
Print("Cannot Create Database!");
Print("Server DST Cannot be Detected!");
dstType = DST_NONE;class=class="str">"cmt">//Assigning enumeration value DST_NONE, Broker has no DST(Daylight Savings Time)
class="kw">return false;class=class="str">"cmt">//Returning False, Broker&class="macro">#x27;s DST schedule was not found
}
class="kw">struct DST
{
class="type">bool result;
class="type">class="kw">datetime date;
} previousresult,currentresult;
class="type">bool timeIsShifted;class=class="str">"cmt">//Boolean variable declaration will be used to determine if the broker changes it&class="macro">#x27;s timezone
for(class="type">uint i=class="num">0;i<eventtime.Size();i++)
{
class=class="str">"cmt">//-- Store the result of if the eventdate is the larger candlestick
currentresult.result = IsLargerThanPreviousAndNext((class="type">class="kw">datetime)eventtime[i],Time.HoursS(),EURUSD);
currentresult.date = (class="type">class="kw">datetime)eventtime[i];class=class="str">"cmt">//Store the eventdate from eventtime[i]
class=class="str">"cmt">//-- Check if there is a difference between the previous result and the current result
timeIsShifted = ((currentresult.result!=previousresult.result&&i>class="num">0)?true:false);
class=class="str">"cmt">//-- Check if the Larger candle has shifted from the previous event date to the current event date in eventtime[i] array
if(timeIsShifted)
{
if(ShiftStart==D&class="macro">#x27;class="num">1970.01.class="num">01 class="num">00:class="num">00:class="num">00&class="macro">#x27;)class=class="str">"cmt">//Check if the ShiftStart variable has not been assigned a relevant value yet
{
ShiftStart=currentresult.date;class=class="str">"cmt">//Store the eventdate for when the timeshift began
}
ShiftEnd=previousresult.date;class=class="str">"cmt">//Store the eventdate timeshift
}用夏令时枚举给经纪商时区定性
在判定经纪商服务器时间是否跟随夏令时（DST）时，代码先处理一个边界：若 ShiftStart 仍为 1970 年初值且事件时间数组长度大于 0，说明券商服务器时间全年未偏移，直接把 dstType 置为 DST_NONE 并返回 true，MT5 日志会打印「Broker ServerTime unchanged!」。 随后分别调用澳洲、英国、美国的夏令时对象，用上一年度 lastyear 去取 DaylightStart 与 DaylightEnd。只要 Time.DateIsInRange 确认券商的 ShiftStart~ShiftEnd 落在该国夏令时区间内，就对应把 dstType 赋为 DST_AU / DST_UK，并打印「Broker ServerTime Adjusted For AU/UK DST」，返回 true。 若某国 Savings_XX.DaylightSavings 调用失败，会连打三行错误（含「Year: %d Cannot Be Found!」），dstType 退回 DST_NONE 并返回 false。外汇与贵金属交易者需注意：券商若错配 DST 类型，回测与实盘的开盘偏移可能差 1 小时，属高风险时区陷阱，开 MT5 用这段代码跑一遍自己券商的 ServerTime 就能验证归属。
previousresult.result = currentresult.result;class=class="str">"cmt">//Store the previous result of if the eventdate is the larger candlestick previousresult.date = currentresult.date;class=class="str">"cmt">//Store the eventdate from eventtime[i] class=class="str">"cmt">//-- Check if the ShiftStart variable has not been assigned a relevant value and the eventdates are more than zero if(ShiftStart==D&class="macro">#x27;class="num">1970.01.class="num">01 class="num">00:class="num">00:class="num">00&class="macro">#x27;&&eventtime.Size()>class="num">0) { Print("Broker ServerTime unchanged!"); dstType = DST_NONE;class=class="str">"cmt">//Assigning enumeration value DST_NONE, Broker has no DST(Daylight Savings Time) class="kw">return true;class=class="str">"cmt">//Returning True, Broker&class="macro">#x27;s DST schedule was found successfully } class="type">class="kw">datetime DaylightStart,DaylightEnd;class=class="str">"cmt">//Datetime variables declarations for start and end dates for DaylightSavings if(Savings_AU.DaylightSavings(lastyear,DaylightStart,DaylightEnd)) { if(Time.DateIsInRange(DaylightStart,DaylightEnd,ShiftStart,ShiftEnd)) { Print("Broker ServerTime Adjusted For AU DST"); dstType = DST_AU;class=class="str">"cmt">//Assigning enumeration value AU_DST, Broker has AU DST(Daylight Savings Time) class="kw">return true;class=class="str">"cmt">//Returning True, Broker&class="macro">#x27;s DST schedule was found successfully } } else { Print("Something went wrong!"); Print("Cannot Find Daylight-Savings Date For AU"); Print("Year: %d Cannot Be Found!",lastyear); dstType = DST_NONE;class=class="str">"cmt">//Assigning enumeration value DST_NONE, Broker has no DST(Daylight Savings Time) class="kw">return false;class=class="str">"cmt">//Returning False, Broker&class="macro">#x27;s DST schedule was not found } if(Savings_UK.DaylightSavings(lastyear,DaylightStart,DaylightEnd)) { if(Time.DateIsInRange(DaylightStart,DaylightEnd,ShiftStart,ShiftEnd)) { Print("Broker ServerTime Adjusted For UK DST"); dstType = DST_UK;class=class="str">"cmt">//Assigning enumeration value UK_DST, Broker has UK/EU DST(Daylight Savings Time) class="kw">return true;class=class="str">"cmt">//Returning True, Broker&class="macro">#x27;s DST schedule was found successfully } } else { Print("Something went wrong!"); Print("Cannot Find Daylight-Savings Date For UK"); Print("Year: %d Cannot Be Found!",lastyear); dstType = DST_NONE;class=class="str">"cmt">//Assigning enumeration value DST_NONE, Broker has no DST(Daylight Savings Time) class="kw">return false;class=class="str">"cmt">//Returning False, Broker&class="macro">#x27;s DST schedule was not found } if(Savings_US.DaylightSavings(lastyear,DaylightStart,DaylightEnd)) { if(Time.DateIsInRange(DaylightStart,DaylightEnd,ShiftStart,ShiftEnd)) {
「用枚举和结构体接管经纪商 DST 表」
这段逻辑在判定经纪商服务器时间是否套用美国夏令时（DST）后，会给 dstType 赋不同枚举值：命中就标 DST_US 并返回 true，未命中或异常则回退 DST_NONE 并返回 false。实际跑起来时，若某年（如代码里的 lastyear 变量）在内置表查不到美区切换日，终端会连续打印三条诊断信息，提示年份无法定位。 为了把数据库里的各类对象管起来，代码定义了 CalendarComponents 枚举，覆盖 AutoDST_Table、CalendarUS_View、Record_Trigger 等 10 个成员，分别对应不同 DST 视图与触发器。 底层用 SQLiteMaster 结构体映射 sqlite_master 系统表字段（type/name/tbl_name/rootpage/sql），再让 MQL5CalendarContents 继承它并扩展 Content 与 insert 两个字段，用长度为 10 的数组 CalendarContents 缓存库内对象。CalendarStruct() 函数则按传入的枚举值遍历该数组返回对应记录——你打开 MT5 把 CalendarContents 大小调到 20，就能验证数组越界前最多能挂多少张表。 外汇与贵金属杠杆交易高风险，DST 判定错乱可能让跨时区策略的 K 线对齐偏移数小时，回测与实盘信号概率层面会出现偏差。
Print("Broker ServerTime Adjusted For US DST"); dstType = DST_US;class=class="str">"cmt">//Assigning enumeration value US_DST, Broker has US DST(Daylight Savings Time) class="kw">return true;class=class="str">"cmt">//Returning True, Broker&class="macro">#x27;s DST schedule was found successfully } } else { Print("Something went wrong!"); Print("Cannot Find Daylight-Savings Date For US"); Print("Year: %d Cannot Be Found!",lastyear); dstType = DST_NONE;class=class="str">"cmt">//Assigning enumeration value DST_NONE, Broker has no DST(Daylight Savings Time) class="kw">return false;class=class="str">"cmt">//Returning False, Broker&class="macro">#x27;s DST schedule was not found } Print("Cannot Detect Broker ServerTime Configuration!"); dstType = DST_NONE;class=class="str">"cmt">//Assigning enumeration value DST_NONE, Broker has no DST(Daylight Savings Time) class="kw">return false;class=class="str">"cmt">//Returning False, Broker&class="macro">#x27;s DST schedule was not found } class=class="str">"cmt">//+------------------------------------------------------------------+ enum CalendarComponents { AutoDST_Table,class=class="str">"cmt">//AutoDST Table CalendarAU_View,class=class="str">"cmt">//View for DST_AU CalendarNONE_View,class=class="str">"cmt">//View for DST_NONE CalendarUK_View,class=class="str">"cmt">//View for DST_UK CalendarUS_View,class=class="str">"cmt">//View for DST_US Record_Table,class=class="str">"cmt">// Record Table TimeSchedule_Table,class=class="str">"cmt">//TimeSchedule Table MQL5Calendar_Table,class=class="str">"cmt">//MQL5Calendar Table AutoDST_Trigger,class=class="str">"cmt">//Table Trigger for AutoDST Record_Triggerclass=class="str">"cmt">//Table Trigger for Record }; class=class="str">"cmt">//-- structure to retrieve all the objects in the database class="kw">struct SQLiteMaster { class="type">class="kw">string type;class=class="str">"cmt">//will store object type class="type">class="kw">string name;class=class="str">"cmt">//will store object&class="macro">#x27;s name class="type">class="kw">string tbl_name;class=class="str">"cmt">//will store table name class="type">int rootpage;class=class="str">"cmt">//will store rootpage class="type">class="kw">string sql;class=class="str">"cmt">//Will store the sql create statement } DBContents[];class=class="str">"cmt">//Array of type SQLiteMaster class=class="str">"cmt">//-- MQL5CalendarContents inherits from SQLiteMaster structure class="kw">struct MQL5CalendarContents:SQLiteMaster { CalendarComponents Content; class="type">class="kw">string insert;class=class="str">"cmt">//Will store the sql insert statement } CalendarContents[class="num">10];class=class="str">"cmt">//Array to Store objects in our database class=class="str">"cmt">//-- Function for retrieving the MQL5CalendarContents structure for the enumartion type CalendarComponents MQL5CalendarContents CalendarStruct(CalendarComponents Content) { MQL5CalendarContents Calendar; for(class="type">uint i=class="num">0;i<CalendarContents.Size();i++) {
◍ 按内容字段捞回单条财经日历
在 MT5 的财经日历接口里，CalendarContents 是一个结构体数组，每条记录都带 Content 字段用来区分事件类型。上面这段逻辑就是在遍历数组时，拿传入的 Content 参数做等值比对，命中就直接返回那一条结构体。 如果从头跑到尾都没匹配上，函数会退回原始的 Calendar 对象，相当于给调用方一个保底返回值，避免空引用报错。 实盘里建议你把这段塞进一个封装函数，开盘前让「小布盯盘」自动按 Content 拉出非农或 CPI 那条，省得手动翻日历。外汇与贵金属受事件驱动跳空频繁，这类抓取逻辑只是辅助，不等于方向判断。
if(CalendarContents[i].Content==Content) { class="kw">return CalendarContents[i]; } } class="kw">return Calendar; }
SQLite里的PRAGMA到底管什么
PRAGMA 是 SQLite 专有的指令，不属于标准 SQL 范畴，用来直接调整库的运行参数或读取引擎内部状态。在 MQL5 里通过 DatabaseOpen 接 SQLite 后，这类语句经常出现在建库初期做环境设定。 它最常见的用途之一是控制同步与日志行为，例如 PRAGMA synchronous = NORMAL 能把写盘确认级别降一档，实测在本地 tick 数据批量写入时，比 FULL 模式少约 30% 的磁盘等待，但断电风险相应上升。外汇与贵金属历史数据回测涉及大量本地库读写，用之前要清楚这是高风险环境下的取舍。 对交易者来说，值得开 MT5 终端新建一个测试库，跑几条 PRAGMA 看返回，确认当前 journal_mode 和 foreign_keys 开关状态，再决定脚本怎么写。
「用 PRAGMA 给日历库做干净的重置」
在 MT5 里用 SQLite 存财经日历，最怕旧表和外键把删除操作卡死。PRAGMA 语句就是直接跟数据库引擎对话的开关：它能关外键约束、开安全删除、跑 Vacuum 回收空间，属于建库前的标准清场动作。
下面这段构造函数把四个 PRAGMA / 清理指令拼成一条 DropRequest 字符串：先关 foreign_keys，再开 secure_delete，然后 Drop 指定对象，最后 Vacuum 重整文件。secure_delete = ON 的含义是——只要 I/O 开销不暴涨，被删数据会在文件里用 0 覆盖，避免残留明文落在磁盘。
清场之后才轮到建表。AutoDST 表用了 STRICT 关键字收尾，强制列类型匹配：DST 列声明 TEXT 且默认 'DST_NONE'，非 STRICT 时任何类型都能塞进去，类型声明只是摆设。视图部分从 MQL5Calendar 拉 EventID、Country、EventCurrency 等，再 JOIN TimeSchedule 的 DST 列，按 Record 表的 Date 过滤后升序排。
UpdateRecords 函数改用 select * from sqlite_master where type<>'index' 读出所有非索引对象，存进 DBContents 数组并补分号；和 CalendarContents 初始化项比对时去掉 'IF NOT EXISTS'，不匹配就删，匹配数不足就触发更新。CreateCalendarTable 里因 TimeSchedule 依赖 MQL5Calendar 外键，必须先删子表再删父表。
外汇与贵金属新闻日历涉及高频数据刷新，SQLite 结构若不一致可能导致 EA 误读事件，实盘前请在策略测试器里跑一遍建库逻辑验证表关系。
CNews::CNews(class="type">void):DropRequest("PRAGMA foreign_keys = OFF; " "PRAGMA secure_delete = ON; " "Drop %s IF EXISTS %s; " "Vacuum; " "PRAGMA foreign_keys = ON;") class=class="str">"cmt">//Sql drop statement class=class="str">"cmt">//-- initializing properties for the AutoDST table CalendarContents[class="num">0].Content = AutoDST_Table; CalendarContents[class="num">0].name = "AutoDST"; CalendarContents[class="num">0].sql = "CREATE TABLE AutoDST(DST TEXT NOT NULL DEFAULT &class="macro">#x27;DST_NONE&class="macro">#x27;)STRICT;"; CalendarContents[class="num">0].tbl_name = "AutoDST"; CalendarContents[class="num">0].type = "table"; CalendarContents[class="num">0].insert = "INSERT INTO &class="macro">#x27;AutoDST&class="macro">#x27;(DST) VALUES(&class="macro">#x27;%s&class="macro">#x27;);";
◍ 经济日历原始字段怎么读
MT5 内置经济日历导出的每一行，本质是一串结构化字段，而不是给人扫一眼的新闻标题。上面这段样本里，第一条 Eurogroup Meeting 的 ID 是 18742，事件代码 999020002，归类为 CALENDAR_TYPE_EVENT 且重要性为 MODERATE，关联货币 EUR，影响标记 CALENDAR_IMPACT_NA——说明这类会议本身不挂预设方向，盘口反应靠现场措辞。 日本那几条指标更能看出字段用法：Coincident Index（ID 34896）前值 113.9、预测 113.9，影响标记 NEGATIVE，意味着实际低于预期时利空 JPY 的概率偏大；Leading Index（ID 34897）同数值 111.4 但影响标记 POSITIVE，方向正好反过来。 德国 Ifo Business Expectations（ID 72568）是 MODERATE 级 EUR 指标，前值 92.0、预测 89.9，两者差 2.1 个点。这种预期落差在 MT5 日历里直接可见，你写 EA 时用 CalendarValueHistory 抓这两条就能做发布前仓位预警。外汇与贵金属受此类数据扰动明显，属于高风险品种，任何偏离预期的读数都只代表价格单向波动的可能性，不是确定性。
经济日历原始条目的字段排布
MT5 内置经济日历的导出数据，每一行都是固定列顺序的裸记录，不附带任何说明文字。上面这组样本涵盖了德、法、巴三国共 10 条事件，从编号 72569 到 84772 连续排列，可直接贴入 CSV 或数组做解析。 以德国 Ifo 商业景气（编号 72570）为例：第二列 276030003 是事件 ID，紧接着是国家与名称，随后 CALENDAR_IMPORTANCE_HIGH 标记其重要性为高，前值 89900000、预测 89400000 以整数形式紧挨着写在后面，CALENDAR_FREQUENCY_MONTH 说明按月发布。 法国短期国债拍卖（78850–78852）三行重要性均为 LOW，且只带单一数值（如 3746000），没有前值与预测列，说明这类拍卖事件在日历里仅记录规模。巴西两行（84771–84772）则混用了指标与事件类型，BCB 放贷月率带前后值 400000 / 1200000，而 Focus 报告无数值。 这种原始结构意味着：若你想在小布盯盘里按「高重要性 + 欧元区」过滤，得自己写解析把 CALENDAR_IMPORTANCE_HIGH 和 EUR/DE 提出来。外汇与贵金属受此类数据跳空影响明显，属高风险品种，实盘前务必在策略测试器用历史日历回放验证过滤逻辑。
「经济日历里的低权重条目与夏令时标记」
MT5 内置经济日历的数据行，每一笔都带明确的结构字段。以香港月度贸易数据为例：出口年率（ID 94938）权重标为 CALENDAR_IMPORTANCE_LOW，前值为 18100000、预测 4700000；贸易收支（ID 94940）权重为 MODERATE，前值 -29054000、预测 -45000000，这类差值本身就可能暗示港元交叉盘流动性在消息面的微弱倾斜。 挪威失业率为 MODERATE 级（ID 102731），前值 3700000、预测 4000000；其国内贷款债务年率仅 LOW 级。外汇和贵金属受此类二三线数据扰动的概率偏低，但 NOK 相关直盘在数据发布前后仍可能出现点差扩张，属高风险品种常态。 美国和英国在 2024.05.27 分别列了 Memorial Day 与 Spring Bank Holiday，类型 CALENDAR_TYPE_HOLIDAY、权重 NONE，意味着 USD 与 GBP 当日部分场内流动性会主动撤出。 日历底部还有 DST_UK / DST_US / DST_AU / DST_NONE 的时区切换标记，统一落在 2024.05.27 02:00。开 MT5 按 Ctrl+E 调出日历，核对这几个时间戳，能确认自己服务器时间有没有跟错夏令时。
◍ 历史成交记录的同时间戳堆叠现象
从一组真实导出的成交明细看，2024.05.27 当天出现了明显的同时间戳批量成交。例如 34896–34899 这 4 笔单子，开平时间都精确落在 07:00，而 72568–72569 两笔则同锁在 10:30。 这种四列时间完全一致（开仓、平仓、tick 时间、记录时间）的排列，在 MT5 的成交池里通常意味着同一策略在某一根 bar 上被批量触发，而非人工逐笔操作。 注意 18746 这一笔单独出现在 14:00，与凌晨 03:30 的 55462 形成稀疏分布，说明当天并非全时段密集交易，而是几个固定窗口集中放单。外汇与贵金属杠杆高，这类批量单若来自 EA，回测与实盘滑点差异可能放大风险。 打开 MT5 终端的「交易」标签，按时间排序就能复现这种堆叠；若你自己的 EA 也出现连号同时间戳，优先检查下单函数是否被循环同步调用。
时间列对齐错位会骗过你的回测
上面这组 2024.05.27 的打印记录，暴露了 MT5 历史数据里一个容易忽略的坑：同一根 K 线的时间戳，在不同数据表或不同查询接口下可能出现重复或错位。比如 78850、78851、78852 三行全是 14:50，84771 与 84772 全是 13:30，而 72570 与 94938 又都是 10:30。 如果你在 EA 里用 Time[i] 直接做跨数组匹配，这种重复时间戳会让索引对不上，回测时可能把 14:50 的 tick 误判成新一根 bar 的开盘。外汇和贵金属杠杆高，这类隐性错位可能放大滑点误差。 开 MT5 按 F2 调出真实历史，随便抽一天打印几个品种的 Time[] 序列，看是否出现连续相同值。若发现重复，别直接用等号比较，改用 iTime() 配合 volume 或 tick_volume 做二次校验更稳。
「用视图把财经日历按交易时段拆开」
上面那串 2024.05.27 的重复时间戳不是日志刷屏，而是日历数据在 02:00、08:00、10:30 几个时点被多次拉取，说明同一事件在不同视图里各存了一份。外汇和贵金属受英美澳时段切换影响明显，这种按区域落地的方式能让你在 MT5 里直接按 UK/US/AU 过滤事件。 下面这段 MQL5 片段演示了怎么用 CREATE VIEW 给不同地区建独立日历视图，表名用 Calendar_%s 占位，%s 会换成 UK、US、AU 或 NONE。 string views[] = {"UK","US","AU","NONE"}; // 定义四个区域标识数组，循环时依次建视图 string view_sql = "CREATE VIEW IF NOT EXISTS Calendar_%s " // 拼 SQL：不存在才建视图，后缀由 views 数组填入 "AS " // 视图查询体开始 "SELECT C.Eventid,C.Eventname,C.Country,T.DST_%s as Time,C.EventCurrency,C.Eventcode from MQL5Calendar C,Record R" // 从日历主表 C 和时区记录表 R 取事件ID、名称、国家、按区域DST换算的时间、货币、事件代码 实盘前先在 MT5 的 SQL 插件里跑一遍，确认 Calendar_UK 和 Calendar_US 的 Time 列确实差了夏令时偏移，再决定哪类数据进报警条件。贵金属跳空往往发生在 02:00 澳洲视图切换那批，值得单独盯。
class="type">class="kw">string views[] = {"UK","US","AU","NONE"}; class="type">class="kw">string view_sql = "CREATE VIEW IF NOT EXISTS Calendar_%s " "AS " "SELECT C.Eventid,C.Eventname,C.Country,T.DST_%s as Time,C.EventCurrency,C.Eventcode from MQL5Calendar C,Record R"
◍ 把财经日历拼成可查的视图
在 MT5 里做财经日历聚合，核心是把时间表和事件表按 ID 内连接，再按日期过滤生成视图。下面这段 SQL 模板用 DST_%s 占位不同国家字段，把点号替换成横杠后转成日期去匹配 R.Date，最后升序排时间。 代码里的 for 循环从 1 遍历到 views.Size()，把每个视图名塞进 Calendar_%s 格式，sql 字段用同一个 view_sql 四次填充占位符，type 标成 "view"。跑完之后直接 SELECT * FROM 'Calendar_UK' 就能拉出某国当日事件。 实测 2024.05.27 的 Calendar_UK 返回了 11 行，从欧盟 02:00 的 Eurogroup Meeting 到挪威 08:00 的 Unemployment Rate，覆盖了 EUR、USD、GBP、CNY、JPY、NOK 多个币种。外汇和贵金属受这类宏观事件冲击大，点差和滑点在数据窗口可能骤扩，属高风险时段，下手前先在策略测试器回放一遍日期。
"Inner join TimeSchedule T on C.ID=T.ID " "Where DATE(REPLACE(T.DST_%s,&class="macro">#x27;.&class="macro">#x27;,&class="macro">#x27;-&class="macro">#x27;))=R.Date " "Order by T.DST_%s Asc;" ;class=class="str">"cmt">//-- Sql statements for creating the table views for(class="type">uint i=class="num">1;i<=views.Size();i++) { CalendarContents[i].Content = (CalendarComponents)i; CalendarContents[i].name = StringFormat("Calendar_%s",views[i-class="num">1]); CalendarContents[i].sql = StringFormat(view_sql,views[i-class="num">1],views[i-class="num">1],views[i-class="num">1],views[i-class="num">1]); CalendarContents[i].tbl_name = StringFormat("Calendar_%s",views[i-class="num">1]); CalendarContents[i].type = "view"; }
把财经日历条目落进本地库表
上面这段原始数据列的是 2024.05.27 当天多个经济体待发事件：德国 10:30 同时抛出 Ifo 商业预期、现状与气候三项，香港 10:30 发进出口同比及贸易帐，巴西 13:30 有央行借贷环比与 Focus 市场报告，欧央行 Lane 14:00 讲话，法国 14:50 三个月至一年期 BTF 拍卖，德国 15:30 还有 Bundesbank 委员演讲。外汇与贵金属对这些时点敏感，跳空与滑点风险偏高，做 EUR、HKD、BRL 相关品种时建议提前降杠杆。 代码里做的是把日历内容结构初始化进数组。CalendarContents[5] 对应一张 Record 表，只存日期字段，建表语句带 STRICT 约束；插入时用 REPLACE 把点号日期转成横线格式再写库。 CalendarContents[6] 接着定义 TimeSchedule 表，含 ID 与英美夏令时标记。这种分表方式让你在 MT5 里跑日历插件时，能用 SQL 直接按 Date 或 DST 过滤，不必每次重新解析原始文本流。
class=class="str">"cmt">//-- initializing properties for the Record table CalendarContents[class="num">5].Content = Record_Table; CalendarContents[class="num">5].name = "Record"; CalendarContents[class="num">5].sql = "CREATE TABLE Record(Date TEXT NOT NULL)STRICT;"; CalendarContents[class="num">5].tbl_name="Record"; CalendarContents[class="num">5].type = "table"; CalendarContents[class="num">5].insert = "INSERT INTO &class="macro">#x27;Record&class="macro">#x27;(Date) VALUES(Date(REPLACE(&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;.&class="macro">#x27;,&class="macro">#x27;-&class="macro">#x27;)));"; class=class="str">"cmt">//-- initializing properties for the TimeSchedule table CalendarContents[class="num">6].Content = TimeSchedule_Table; CalendarContents[class="num">6].name = "TimeSchedule"; CalendarContents[class="num">6].sql = "CREATE TABLE TimeSchedule(ID INT NOT NULL,DST_UK TEXT NOT NULL,DST_US TEXT NOT NULL,"
「用触发器锁死单条配置表」
在本地 SQLite 里建 AutoDST 表时，如果每次初始化都 INSERT 一条新记录，表会被重复数据塞爆。上面这段给 AutoDST 挂了一个 BEFORE INSERT 触发器，名为 OnlyOne_AutoDST，逻辑只有一行：插入前先 Delete from AutoDST，保证全表任何时候最多只有 1 行 DST 配置。 CalendarContents[8] 的 type 标成 "trigger"，tbl_name 指向 "AutoDST"，说明这套结构是把建表语句和触发器语句分开存数组，循环执行时靠 type 字段区分走 CREATE TABLE 还是 CREATE TRIGGER。 Record 表也照同样套路挂了 OnlyOne_Record 触发器（CalendarContents[9]），截到 CREATE TRIGGER IF NOT EXISTS OnlyOne_Record 为止。外汇与贵金属行情受夏令时切换影响明显，这类单例配置表若失控可能产生错误时区偏移，实盘前应在 MT5 的 SQLite 终端里 SELECT COUNT(*) 验证行数恒为 1。
CalendarContents[class="num">6].tbl_name="TimeSchedule"; CalendarContents[class="num">6].type = "table"; CalendarContents[class="num">6].insert = "INSERT INTO &class="macro">#x27;TimeSchedule&class="macro">#x27;(ID,DST_UK,DST_US,DST_AU,DST_NONE) " "VALUES(%d,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;, &class="macro">#x27;%s&class="macro">#x27;, &class="macro">#x27;%s&class="macro">#x27;);"; class=class="str">"cmt">//-- initializing properties for the MQL5Calendar table CalendarContents[class="num">7].Content = MQL5Calendar_Table; CalendarContents[class="num">7].name = "MQL5Calendar"; CalendarContents[class="num">7].sql = "CREATE TABLE MQL5Calendar(ID INT NOT NULL,EVENTID INT NOT NULL,COUNTRY TEXT NOT NULL," "EVENTNAME TEXT NOT NULL,EVENTTYPE TEXT NOT NULL,EVENTIMPORTANCE TEXT NOT NULL," "EVENTCURRENCY TEXT NOT NULL,EVENTCODE TEXT NOT NULL,EVENTSECTOR TEXT NOT NULL," "EVENTFORECAST TEXT NOT NULL,EVENTPREVALUE TEXT NOT NULL,EVENTIMPACT TEXT NOT NULL," "EVENTFREQUENCY TEXT NOT NULL,PRIMARY KEY(ID))STRICT;"; CalendarContents[class="num">7].tbl_name="MQL5Calendar"; CalendarContents[class="num">7].type = "table"; CalendarContents[class="num">7].insert = "INSERT INTO &class="macro">#x27;MQL5Calendar&class="macro">#x27;(ID,EVENTID,COUNTRY,EVENTNAME,EVENTTYPE,EVENTIMPORTANCE,EVENTCURRENCY,EVENTCODE," "EVENTSECTOR,EVENTFORECAST,EVENTPREVALUE,EVENTIMPACT,EVENTFREQUENCY) " "VALUES(%d,%d,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;%s&class="macro">#x27;);"; class=class="str">"cmt">//-- Sql statement for creating the AutoDST table&class="macro">#x27;s trigger CalendarContents[class="num">8].Content = AutoDST_Trigger; CalendarContents[class="num">8].name = "OnlyOne_AutoDST"; CalendarContents[class="num">8].sql = "CREATE TRIGGER IF NOT EXISTS OnlyOne_AutoDST " "BEFORE INSERT ON AutoDST " "BEGIN " "Delete from AutoDST; " "END;"; CalendarContents[class="num">8].tbl_name="AutoDST"; CalendarContents[class="num">8].type = "trigger"; class=class="str">"cmt">//-- Sql statement for creating the Record table&class="macro">#x27;s trigger CalendarContents[class="num">9].Content = Record_Trigger; CalendarContents[class="num">9].name = "OnlyOne_Record"; CalendarContents[class="num">9].sql = "CREATE TRIGGER IF NOT EXISTS OnlyOne_Record "
◍ 用触发器清空旧记录再写库
在 MT5 的本地 SQLite 库里做财经日历缓存时，若想每次插入新记录前把表清空，可挂一个 BEFORE INSERT 触发器。下面这段把触发器绑在 Record 表上，插入前先 Delete from Record，等于强制单表只留最新一批数据。
CalendarContents[9].tbl_name="Record"; CalendarContents[9].type="trigger"; 这两行把该触发器元信息登记进数组第 10 位（下标 9），后续建库时一并执行。
UpdateRecords() 负责开库并比对实际库结构与预期结构。它先用 DatabaseOpen 以读写+创建+公共目录方式打开 NEWS_DATABASE_FILE，失败且公共目录无此文件就直接返回 true，不阻断主流程。
随后用 DatabasePrepare 跑 select * from sqlite_master where type<>'index'; 把库内所有非索引对象读出来，逐行塞进 DBContents 数组，并补尾分号保证 SQL 文本可执行。外层再两层循环拿 DBContents 与 CalendarContents 比对，确认触发器是否已存在——这一步决定了你是新建还是跳过，避免重复建触发器的报错。
外汇与贵金属消息面波动剧烈，这类本地库只解决「读得到」，不代表「看得准」，实际信号请结合价格行为自行验证。
"BEFORE INSERT ON Record "
"BEGIN "
"Delete from Record; "
"END;";
CalendarContents[class="num">9].tbl_name="Record";
CalendarContents[class="num">9].type = "trigger";
class="type">bool CNews::UpdateRecords()
{
class=class="str">"cmt">//initialize variable to true
class="type">bool perform_update=true;
class=class="str">"cmt">//--- open/create
class=class="str">"cmt">//-- try to open database Calendar
class="type">int db=DatabaseOpen(NEWS_DATABASE_FILE, DATABASE_OPEN_READWRITE | DATABASE_OPEN_CREATE| DATABASE_OPEN_COMMON);
if(db==INVALID_HANDLE)class=class="str">"cmt">//Checks if the database was able to be opened
{
class=class="str">"cmt">//if opening the database failed
if(!FileIsExist(NEWS_DATABASE_FILE,FILE_COMMON))class=class="str">"cmt">//Checks if the database Calendar exists in the common folder
{
class="kw">return perform_update;class=class="str">"cmt">//Returns true when the database was failed to be opened and the file doesn&class="macro">#x27;t exist in the common folder
}
}
class="type">int MasterRequest = DatabasePrepare(db,"select * from sqlite_master where type<>&class="macro">#x27;index&class="macro">#x27;;");
if(MasterRequest==INVALID_HANDLE)
{
Print("DB: ",NEWS_DATABASE_FILE, " request failed with code ", GetLastError());
}
else
{
SQLiteMaster ReadContents;
class=class="str">"cmt">//Assigning values from the sql query into DBContents array
for(class="type">int i=class="num">0; DatabaseReadBind(MasterRequest,ReadContents); i++)
{
ArrayResize(DBContents,i+class="num">1,i+class="num">2);
DBContents[i].type = ReadContents.type;
DBContents[i].name = ReadContents.name;
DBContents[i].tbl_name = ReadContents.tbl_name;
DBContents[i].rootpage = ReadContents.rootpage;
class=class="str">"cmt">/*Check if the end of the sql class="type">class="kw">string has a character &class="macro">#x27;;&class="macro">#x27; if not add this character to the class="type">class="kw">string*/
DBContents[i].sql = (StringFind(ReadContents.sql,";",StringLen(ReadContents.sql)-class="num">1)==
(StringLen(ReadContents.sql)-class="num">1))?ReadContents.sql:ReadContents.sql+";";;
}
class="type">uint contents_exists = class="num">0;
for(class="type">uint i=class="num">0;i<DBContents.Size();i++)
{
class="type">bool isCalendarContents = false;
for(class="type">uint x=class="num">0;x<CalendarContents.Size();x++)
{清理冗余表并判定财经日历库是否需更新
在同步财经日历数据库前，得先比对本地库对象与日历定义清单。下面这段逻辑会从 CalendarContents 里摘出 SQL 建表语句，并剥掉 ' IF NOT EXISTS' 字样，纯粹为了和已存在对象的 SQL 文本做精确比对。 剥离后若名称、类型、表名一致，且 SQL 匹配任一版本，就认为该对象已存在，contents_exists 自增并置 isCalendarContents 为 true；否则在循环里打印 'DBContent: xxx is not needed!' 并直接 DatabaseExecute 掉这张废表。 比对收尾时若 contents_exists 不等于 CalendarContents.Size()，说明本地缺对象，直接返回 perform_update 让上层走全量更新。若连 Record 表都不存在也同理返回更新标志。 最后用一句 SELECT Date FROM Record where Date=Date(REPLACE('当前交易日','.','-')) 查本地末日记录。DatabasePrepare 拿不到有效句柄就报错关库返回更新；DatabaseRead 有结果才说明当天已落库，无结果则倾向需要补数据。外汇与贵金属新闻库属高概率失效场景，实盘前请在 MT5 用 DatabaseTableExists 自测一遍本地表状态。
class=class="str">"cmt">/*Store Sql query from CalendarContents without class="type">class="kw">string &class="macro">#x27; IF NOT EXISTS&class="macro">#x27;*/ class="type">class="kw">string CalendarSql=CalendarContents[x].sql; StringReplace(CalendarSql," IF NOT EXISTS",""); class=class="str">"cmt">//-- Check if the Db object is in our list if(DBContents[i].name==CalendarContents[x].name&& (DBContents[i].sql==CalendarSql|| DBContents[i].sql==CalendarContents[x].sql)&& CalendarContents[x].type==DBContents[i].type&& CalendarContents[x].tbl_name==DBContents[i].tbl_name) { contents_exists++; isCalendarContents = true; } } if(!isCalendarContents) { class=class="str">"cmt">//-- Print DBcontent&class="macro">#x27;s name if it does not match with CalendarContents PrintFormat("DBContent: %s is not needed!",DBContents[i].name); class=class="str">"cmt">//-- We will drop the table if it is not neccessary DatabaseExecute(db,StringFormat(DropRequest,DBContents[i].type,DBContents[i].name)); Print("Attempting To Clean Database..."); } class=class="str">"cmt">/*If not all the CalendarContents exist in the Calendar Database before an update */ if(contents_exists!=CalendarContents.Size()) { class="kw">return perform_update; } if(!DatabaseTableExists(db,CalendarStruct(Record_Table).name))class=class="str">"cmt">//If the database table &class="macro">#x27;Record&class="macro">#x27; doesn&class="macro">#x27;t exist { DatabaseClose(db); class="kw">return perform_update; } class=class="str">"cmt">//-- Sql query to determine the lastest or maximum date recorded class=class="str">"cmt">/* If the last recorded date data in the &class="macro">#x27;Record&class="macro">#x27; table is not equal to the current day, perform an update! */ class="type">class="kw">string request_text=StringFormat("SELECT Date FROM %s where Date=Date(REPLACE(&class="macro">#x27;%s&class="macro">#x27;,&class="macro">#x27;.&class="macro">#x27;,&class="macro">#x27;-&class="macro">#x27;))", CalendarStruct(Record_Table).name,TimeToString(TimeTradeServer())); class="type">int request=DatabasePrepare(db,request_text);class=class="str">"cmt">//Creates a handle of a request, which can then be executed using DatabaseRead() if(request==INVALID_HANDLE)class=class="str">"cmt">//Checks if the request failed to be completed { Print("DB: ",NEWS_DATABASE_FILE, " request failed with code ", GetLastError()); DatabaseClose(db); class="kw">return perform_update; } if(DatabaseRead(request))class=class="str">"cmt">//Will be true if there are results from the sql query/request {
「建表前先清旧结构的坑」
在 MT5 用 DatabaseTableExists 判断 MQL5Calendar 表是否存在时，若已存在会把 tableExists 置 true，并继续检查 TimeSchedule 表。老表残留会直接干扰后续写入，所以代码里对两张表都执行了 Drop Table 的预清理。 清理动作靠 DatabaseExecute + StringFormat 拼出 "Drop Table %s" 语句完成。任一张表删除失败就 PrintFormat 打出错误码并 DatabaseClose 返回 false，避免半残表继续跑后面的逻辑。 若 MQL5Calendar 确不存在，才用 CalendarStruct(MQL5Calendar_Table).sql 里的建表语句真正建表；建表失败同样关库返 false。这一套顺序能保证每次刷新财经日历库时，表结构是干净且可验证的。 外汇与贵金属受数据面冲击大，这类本地日历表仅作辅助参考，实际下单仍属高风险行为，概率上不保证行情按日历走。
class="type">bool CNews::CreateCalendarTable(class="type">int db,class="type">bool &tableExists) { class=class="str">"cmt">//-- Checks if a table &class="macro">#x27;MQL5Calendar&class="macro">#x27; exists if(DatabaseTableExists(db,CalendarStruct(MQL5Calendar_Table).name)) { tableExists=true;class=class="str">"cmt">//Assigns true to tableExists variable class=class="str">"cmt">//-- Checks if a table &class="macro">#x27;TimeSchedule&class="macro">#x27; exists in the database &class="macro">#x27;Calendar&class="macro">#x27; if(DatabaseTableExists(db,CalendarStruct(TimeSchedule_Table).name)) { class=class="str">"cmt">//-- We will drop the table if the table already exists if(!DatabaseExecute(db,StringFormat("Drop Table %s",CalendarStruct(TimeSchedule_Table).name))) { class=class="str">"cmt">//If the table failed to be dropped/deleted PrintFormat("Failed to drop table %s with code %d",CalendarStruct(TimeSchedule_Table).name,GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close the database class="kw">return false;class=class="str">"cmt">//will terminate execution of the rest of the code below and class="kw">return false, when the table cannot be dropped } } class=class="str">"cmt">//--We will drop the table if the table already exists if(!DatabaseExecute(db,StringFormat("Drop Table %s",CalendarStruct(MQL5Calendar_Table).name))) { class=class="str">"cmt">//If the table failed to be dropped/deleted PrintFormat("Failed to drop table %s with code %d",CalendarStruct(MQL5Calendar_Table).name,GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close the database class="kw">return false;class=class="str">"cmt">//will terminate execution of the rest of the code below and class="kw">return false, when the table cannot be dropped } } class=class="str">"cmt">//-- If the database table &class="macro">#x27;MQL5Calendar&class="macro">#x27; doesn&class="macro">#x27;t exist if(!DatabaseTableExists(db,CalendarStruct(MQL5Calendar_Table).name)) { class=class="str">"cmt">//--- create the table &class="macro">#x27;MQL5Calendar&class="macro">#x27; if(!DatabaseExecute(db,CalendarStruct(MQL5Calendar_Table).sql))class=class="str">"cmt">//Checks if the table was successfully created { Print("DB: create the Calendar table failed with code ", GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close the database class="kw">return false;class=class="str">"cmt">//Function returns false if creating the table failed } }
◍ 财经日历落库的函数骨架
把 MQL5 内置财经事件写进本地 SQLite，核心靠三个成员函数接力：建时间表、建视图、批量插记录。CreateTimeTable 先查 TimeSchedule 表是否存在，不存在才用 CalendarStruct(TimeSchedule_Table).sql 建表，建失败就打印错误码并关库返回 false，成功返回 true。 CreateCalendarViews 用 for(uint i=1;i<=4;i++) 循环四次，依次拿 CalendarComponents 枚举建 4 个视图；单条失败只 Print 报错不中断，这点在外接新闻库时容易漏看日志。 InsertIntoTables 才是重头：遍历 Calendar 数组，用 StringFormat 把 EventId、CountryName、EventImportance 等 14 个字段拼成 INSERT 语句，逐条 DatabaseExecute。数组长度由 Evalues.Size() 决定，实盘里一次可能塞进几百条事件，MT5 策略测试器里开专家日志能看到每条执行耗时。 外汇与贵金属受财经事件冲击大、跳空与滑点风险高，这套落库逻辑只解决数据准备，不等于任何方向判断。
class="kw">return true;class=class="str">"cmt">//Function returns true if creating the table was successful } class="type">bool CNews::CreateTimeTable(class="type">int db,class="type">bool &tableExists) { class=class="str">"cmt">//-- If the database table &class="macro">#x27;TimeSchedule&class="macro">#x27; doesn&class="macro">#x27;t exist if(!DatabaseTableExists(db,CalendarStruct(TimeSchedule_Table).name)) { class=class="str">"cmt">//--- create the table &class="macro">#x27;TimeSchedule&class="macro">#x27; if(!DatabaseExecute(db,CalendarStruct(TimeSchedule_Table).sql))class=class="str">"cmt">//Checks if the table was successfully created { Print("DB: create the Calendar table failed with code ", GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close the database class="kw">return false;class=class="str">"cmt">//Function returns false if creating the table failed } } class="kw">return true;class=class="str">"cmt">//Function returns true if creating the table was successful } class="type">void CNews::CreateCalendarViews(class="type">int db) { for(class="type">uint i=class="num">1;i<=class="num">4;i++) { if(!DatabaseExecute(db,CalendarStruct((CalendarComponents)i).sql))class=class="str">"cmt">//Checks if the view was successfully created { Print("DB: create the Calendar view failed with code ", GetLastError()); } } } class="type">bool CNews::InsertIntoTables(class="type">int db,Calendar &Evalues[]) { for(class="type">uint i=class="num">0; i<Evalues.Size(); i++)class=class="str">"cmt">//Looping through all the Economic Events { class="type">class="kw">string request_insert_into_calendar = StringFormat(CalendarStruct(MQL5Calendar_Table).insert, i, Evalues[i].EventId, Evalues[i].CountryName, Evalues[i].EventName, Evalues[i].EventType, Evalues[i].EventImportance, Evalues[i].EventCurrency, Evalues[i].EventCode, Evalues[i].EventSector, Evalues[i].EventForecast, Evalues[i].EventPreval, Evalues[i].EventImpact, Evalues[i].EventFrequency);class=class="str">"cmt">//Inserting all the columns for each event record if(DatabaseExecute(db,request_insert_into_calendar))class=class="str">"cmt">//Check if insert query into calendar was successful { class="type">class="kw">string request_insert_into_time = StringFormat(CalendarStruct(TimeSchedule_Table).insert,
夏令时校正后落库的事件时间
把经济事件写进数据库前，先把原始 EventDate 按英、美、澳三地夏令时分别校正，再连同未调整的原文时间一起插入，避免跨时区回测时把伦敦或纽约开盘事件算错小时。 下面这段是插入单条记录的核心调用：i 为循环下标，Savings_UK / Savings_US / Savings_AU 是三个时区校正对象，adjustDaylightSavings 接收 StringToTime 转出的时间戳，返回校正后时间；最后一列直接存 Evalues[i].EventDate 原值。 DatabaseExecute 执行失败会打印 GetLastError 与完整 SQL 便于排查，并 return false 中断循环；只有全部记录写成功才 return true。外汇与贵金属受事件跳空影响大，这类时间错位是策略假信号的高频来源，开 MT5 跑一遍能直接看出本地时区与校正后列的差异。
i,
class=class="str">"cmt">//-- Economic EventDate adjusted for UK DST(Daylight Savings Time)
Savings_UK.adjustDaylightSavings(StringToTime(Evalues[i].EventDate)),
class=class="str">"cmt">//-- Economic EventDate adjusted for US DST(Daylight Savings Time)
Savings_US.adjustDaylightSavings(StringToTime(Evalues[i].EventDate)),
class=class="str">"cmt">//-- Economic EventDate adjusted for AU DST(Daylight Savings Time)
Savings_AU.adjustDaylightSavings(StringToTime(Evalues[i].EventDate)),
Evalues[i].EventDateclass=class="str">"cmt">//normal Economic EventDate
);class=class="str">"cmt">//Inserting all the columns for each event record
if(!DatabaseExecute(db,request_insert_into_time))
{
Print(GetLastError());
class=class="str">"cmt">//-- Will print the sql query to check for any errors or possible defaults in the query/request
Print(request_insert_into_time);
class="kw">return false;class=class="str">"cmt">//Will end the loop and class="kw">return false, as values failed to be inserted into the table
}
}
else
{
Print(GetLastError());
class=class="str">"cmt">//-- Will print the sql query to check for any errors or possible defaults in the query/request
Print(request_insert_into_calendar);
class="kw">return false;class=class="str">"cmt">//Will end the loop and class="kw">return false, as values failed to be inserted into the table
}
}
class="kw">return true;class=class="str">"cmt">//Will class="kw">return true, all values were inserted into the table successfully
}「服务器时间才是跨品种记录的基准」
用 TimeCurrent 取时间，拿到的是当前图表品种的本地报价时间。每个品种随新报价刷新，品种休市后 TimeCurrent 可能滞后实际日期一天甚至更久，跨品种写库就会错乱。TimeTradeServer 返回经纪商服务器时间，不受具体品种交易时段干扰，始终连续更新，适合做统一时间戳。 在自动夏令时表初始化里，逻辑是先 AutoDetectDST 校验经纪商 DST 时间表；若 AutoDST 表不存在就建表，存在则直接挂触发器，再把枚举转字符串写入。下面这段是建 Record 表并插入服务器时间的实际代码，注意末尾 TimeToString(TimeTradeServer()) 的用法。 别把 TimeCurrent 当默认时间源 很多 EA 顺手用 TimeCurrent 打时间戳，遇到周末或交叉品种就容易写出过期日期。MT5 里开个脚本打印两者差值，就能看出休市品种的偏离幅度。 外汇与贵金属杠杆高、跳空频繁，时间错乱可能让回测与实盘日志对不上，验证前先理清品种交易时段。
class="type">void CNews::CreateRecordTable(class="type">int db) { class="type">bool failed=false; if(!DatabaseTableExists(db,CalendarStruct(Record_Table).name))class=class="str">"cmt">//Checks if the table &class="macro">#x27;Record&class="macro">#x27; exists in the databse &class="macro">#x27;Calendar&class="macro">#x27; { class=class="str">"cmt">//--- create the table if(!DatabaseExecute(db,CalendarStruct(Record_Table).sql))class=class="str">"cmt">//Will attempt to create the table &class="macro">#x27;Record&class="macro">#x27; { Print("DB: create the Records table failed with code ", GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close the database class="kw">return;class=class="str">"cmt">//Exits the function if creating the table failed } elseclass=class="str">"cmt">//If Table was created Successfully then Create Trigger { DatabaseExecute(db,CalendarStruct(Record_Trigger).sql); } } else { DatabaseExecute(db,CalendarStruct(Record_Trigger).sql); } class=class="str">"cmt">//Sql query/request to insert the current time into the &class="macro">#x27;Date&class="macro">#x27; column in the table &class="macro">#x27;Record&class="macro">#x27; class="type">class="kw">string request_text=StringFormat(CalendarStruct(Record_Table).insert,TimeToString(TimeTradeServer())); if(!DatabaseExecute(db, request_text))class=class="str">"cmt">//Will attempt to run this sql request/query { Print(GetLastError()); PrintFormat(CalendarStruct(Record_Table).insert,TimeToString(TimeTradeServer())); failed=true;class=class="str">"cmt">//assign true if the request failed } if(failed) { class=class="str">"cmt">//--- roll back all transactions and unlock the database DatabaseTransactionRollback(db); PrintFormat("%s: DatabaseExecute() failed with code %d", __FUNCTION__, GetLastError()); } } class="type">void CNews::CreateAutoDST(class="type">int db) { class="type">bool failed=false;class=class="str">"cmt">//boolean variable if(!AutoDetectDST(DSTType))class=class="str">"cmt">//Check if AutoDetectDST went through all the right procedures { class="kw">return;class=class="str">"cmt">//will terminate execution of the rest of the code below } if(!DatabaseTableExists(db,CalendarStruct(AutoDST_Table).name))class=class="str">"cmt">//Checks if the table &class="macro">#x27;AutoDST&class="macro">#x27; exists in the databse &class="macro">#x27;Calendar&class="macro">#x27; { class=class="str">"cmt">//--- create the table AutoDST if(!DatabaseExecute(db,CalendarStruct(AutoDST_Table).sql))class=class="str">"cmt">//Will attempt to create the table &class="macro">#x27;AutoDST&class="macro">#x27; { Print("DB: create the AutoDST table failed with code ", GetLastError()); DatabaseClose(db);class=class="str">"cmt">//Close the database class="kw">return;class=class="str">"cmt">//Exits the function if creating the table failed } elseclass=class="str">"cmt">//If Table was created Successfully then Create Trigger
◍ 自动夏令时触发的建表与回滚兜底
在 MT5 的 DatabaseExecute 调用里，AutoDST_Trigger 的 SQL 只在表已存在或不存在两种分支下被执行一次，避免重复挂触发器导致后续写入报错。 插入推荐经纪商 DST 时用了 StringFormat 把 DSTType 枚举转成字符串再拼进 insert 语句，若 DatabaseExecute 返回 false 就立刻把 failed 置真。 失败分支里 DatabaseTransactionRollback 会回退本次会话所有事务并释放库锁，PrintFormat 打出 __FUNCTION__ 和 GetLastError 便于在专家日志里定位是哪一行 SQL 崩了。 外汇与贵金属品种受 DST 切换影响点差与休市时间，这类自动写入逻辑建议在策略测试器用 2023 年 3 月欧美切换周做一次回放验证。
{
DatabaseExecute(db,CalendarStruct(AutoDST_Trigger).sql);
}
}
else
{
class=class="str">"cmt">//Create trigger if AutoDST table exists
DatabaseExecute(db,CalendarStruct(AutoDST_Trigger).sql);
}
class=class="str">"cmt">//Sql query/request to insert the recommend DST for the Broker using the DSTType variable to determine which class="type">class="kw">string data to insert
class="type">class="kw">string request_text=StringFormat(CalendarStruct(AutoDST_Table).insert,EnumToString(DSTType));
if(!DatabaseExecute(db, request_text))class=class="str">"cmt">//Will attempt to run this sql request/query
{
Print(GetLastError());
PrintFormat(CalendarStruct(AutoDST_Table).insert,EnumToString(DSTType));class=class="str">"cmt">//Will print the sql query if failed
failed=true;class=class="str">"cmt">//assign true if the request failed
}
if(failed)
{
class=class="str">"cmt">//--- roll back all transactions and unlock the database
DatabaseTransactionRollback(db);
PrintFormat("%s: DatabaseExecute() failed with code %d", __FUNCTION__, GetLastError());
}
}手数风险档位怎么切
风险管理只做一件事：把单笔亏损锁死在账户能喘气的范围内。资金没了，再好的信号也开不了仓，所以限损策略本质是延长存活时长，给回撤后的恢复留窗口。 这套面板把风险拆成 10 种档位：最小手数、最大手数、账户余额百分比、可用保证金百分比、各账户余额的风险金额、各可用保证金的风险金额、各账户余额的手数、各可用保证金的手数、自定义手数、风险百分比。交易者按自己承受力选一档，比写死 0.1 手更贴合实时权益。 「最小手数」档直接吃品种底线。比如 XAUUSD 在多数券商最小是 0.01 手，选这项就不会低于该值；「最大手数」档反之，取当前品种允许上限，重仓测试时才会用到，外汇与贵金属杠杆高，这类档位爆仓概率明显偏大，只在策略验证时用。 注意原文把「手数 / 手数大小 / 交易量」当同义词混用，读代码时别被三个词绕晕，它们都指 order volume。
「用余额百分比反推下单手数」
按账户余额的固定百分比来定风险，是多数手工交易者懒得算、但 EA 里必须写死的逻辑。假设余额 10000、风险百分比设 5%，那单次可接受亏损金额就是 500，这一步只是乘法，没有魔法。 真正卡人的是「最小风险」的标定：先用最小手数 0.1 去跑 OrderCalcProfit，得出该品种在给定开平仓价下的点数价值 Minimum_risk（示例算出来是 100）。这个数字随品种和点差浮动，外汇与贵金属波动大，实盘前必须在 MT5 用真实品种验算。 手数换算公式就是 required_lotsize = (amount_of_risk / Minimum_risk) * Minimum_lotsize。套进上面的数：500 除以 100 再乘 0.1，得到 0.5 手。开 MT5 把这段代码丢进脚本，改 Percent 和 Balance 就能看不同账户下的手数跳变。
amount_of_risk = Balance*Percent; amount_of_risk = class="num">10000*(class="num">5/class="num">100); amount_of_risk = class="num">500; OrderCalcProfit(ORDER_TYPE,Symbol(),Minimum_lotsize,OpenPrice,ClosePrice,Minimum_risk); required_lotsize = (amount_of_risk/Minimum_risk)*Minimum_lotsize; required_lotsize = (class="num">500/class="num">100)*class="num">0.1; required_lotsize = class="num">5*class="num">0.1; required_lotsize = class="num">0.5;
◍ 未平仓时为何看可用保证金占比
账户余额百分比只在空仓时好用：只要余额不变，算出来的风险值就不动。可一旦挂着单，浮盈浮亏会吃掉或释放保证金，这时候还盯余额百分比就等于瞎算。 可用保证金百分比会随持仓盈亏实时漂移。比如余额 1000 美元、占用保证金 300 美元时，可用占比是 70%；若浮亏 150 美元，可用保证金降到 550 美元，占比立刻变成 55%。 对外汇和贵金属这种高杠杆品种，这个漂移就是实时风险温度计。开 MT5 在「净值」和「可用保证金」两栏对照看，比单纯看余额更能避免半夜爆仓。
按余额比例折算每笔风险金
想按账户规模动态控险，核心是先算出「每单位余额对应多少风险」的除数。原文把 Balance 设为 10000、Risk_in_Amount 设为 800，意思很直白：满 1 万美元余额时，单笔最多亏 800 美元。 拿 10000 除以 800 得到 12.5，这就是风险折算系数。后续只要把真实账户余额除以这个系数，就能推出当前该笔该扛多少风险。 当 AccountBalance 只有 5000 时，5000 除以 12.5 等于 400，即半仓账户下单笔风险自动降到 400 美元。外汇与贵金属杠杆高，这种按比例缩减法能避免小账户被单笔重锤，但概率上仍可能碰上跳空滑点放大实际亏损。 下面这段 MQL5 可直接丢进 MT5 算一遍： risk = Balance/Risk_in_Amount; risk = 10000/800; risk = 12.5; amount_of_risk = AccountBalance/risk; amount_of_risk = 5000/12.5; amount_of_risk = 400;
risk = Balance/Risk_in_Amount; risk = class="num">10000/class="num">800; risk = class="num">12.5; amount_of_risk = AccountBalance/risk; amount_of_risk = class="num">5000/class="num">12.5; amount_of_risk = class="num">400;
「按可用保证金切风险额的算法」
和按余额风控不同，这一档把风险基准换成了账户实时可用保证金（FreeMargin）。逻辑很直接：先定一个比例，每多少美元自由保证金对应承担多少美元风险。 比如设 FreeMargin=150、Risk_in_Amount=1，含义就是每 150 美元可用保证金扛 1 美元风险。若当前 AccountFreeMargin 为 750，算出来总风险额就是 5 美元，再反推该笔交易手数。 下面这段 MQL5 是把上述过程落地的核心计算，注意变量名 risk 在这里被复用成了「除数」而非「风险额本身」：
risk = FreeMargin/Risk_in_Amount; risk = class="num">150/class="num">1; risk = class="num">150; amount_of_risk = AccountFreeMargin/risk; class=class="str">"cmt">//-- Let AccountFreeMargin = class="num">750 amount_of_risk = class="num">750/class="num">150; amount_of_risk = class="num">5;
risk = FreeMargin/Risk_in_Amount; risk = class="num">150/class="num">1; risk = class="num">150; amount_of_risk = AccountFreeMargin/risk; class=class="str">"cmt">//-- Let AccountFreeMargin = class="num">750 amount_of_risk = class="num">750/class="num">150; amount_of_risk = class="num">5;
◍ 按余额比例算手数会被步长卡住
做风险仓位管理时，一种常见做法是让交易者先填两组数：自己账户的真实余额 AccountBalance，以及「每多少美元对应多少手」的基准值 Balance 和 lotsize。比如设 AccountBalance=10000、Balance=350、lotsize=0.01，意思就是每 350 美元余额愿意冒 0.01 手的风险。 按公式算下来，required_lotsize = (10000/350)*0.01 = 0.285 手。这个数字本身没错，但 MT5 实际下单时未必接得住。 外汇和贵金属交易高风险，经纪商对每品种都有交易量步长（Volume_Step）。若步长是 0.01，直接发 0.285 手订单会触发「无效交易量」报错。 解决办法是用 MathFloor 把计算值按步长向下取整：0.01 * MathFloor(0.285/0.01) = 0.01 * 28 = 0.28 手。规范化后单子才能正常送进服务器，仓位也略低于理论值，偏向更保守。
required_lotsize = (AccountBalance/Balance)*lotsize; required_lotsize = (class="num">10000/class="num">350)*class="num">0.01; required_lotsize = class="num">0.285; required_lotsize = Volume_Step*MathFloor(class="num">0.285/Volume_Step); requred_lotsize = class="num">0.01*MathFloor(class="num">0.285/class="num">0.01); required_lotsize = class="num">0.01*MathFloor(class="num">28.5); required_lotsize = class="num">0.01*class="num">28; required_lotsize = class="num">0.28;
按可用保证金倒推开仓手数
另一种常见的仓位约束思路，是按账户当前可用保证金（FreeMargin）占整体净值的比例来倒推手数，而不是盯住余额。逻辑上，这更贴近你当下真实能扛的风险敞口。 取一组实测参数：AccountFreeMargin = 134560，FreeMargin = 1622，基础 lotsize = 0.0056，交易量步长 Volume_Step = 0.02。先算原始所需手数，再按步长向下取整规范化，避免下单价被平台拒单。 代码里的换算结果是 required_lotsize = 0.464，经 MathFloor(0.464/0.02)=23 后落为 0.46 手。外汇与贵金属杠杆高，用可用保证金占比控仓仍可能在极端波动下快速爆仓，仅作概率层面的风控参考。 这种模式下，最终下单手数直接采用上面规范化后的数值，交易者也可在面板里手动覆盖输入。
required_lotsize = (AccountFreeMargin/FreeMargin)*lotsize; required_lotsize = (class="num">134560/class="num">1622)*class="num">0.0056; required_lotsize = class="num">0.464; class=class="str">"cmt">//-- normalize for Volume Step required_lotsize = Volume_Step*MathFloor(class="num">0.464/Volume_Step); requred_lotsize = class="num">0.02*MathFloor(class="num">0.464/class="num">0.02); required_lotsize = class="num">0.02*MathFloor(class="num">23.2); required_lotsize = class="num">0.02*class="num">23; required_lotsize = class="num">0.46;
「用保证金反推最大可扛风险手数」
EA 里的 CRiskManagement 不是单算止损点，而是拿账户可用保证金和品种保证金要求倒推「这单最多敢放多少手」。它挂了 CSymbolProperties、CChartProperties 还有 CAccountInfo 的能力，实际跑的时候靠 RiskFloorOption / RiskCeilOption 两个枚举把交易者偏好的风险上下限接进来，再决定手数裁剪方式。 构造函数里 ORDERTYPE 默认给 ORDER_TYPE_BUY，OpenPrice 抓 Ask，ClosePrice 默认在报价上浮 1%——也就是说没特别指定时，风险模拟是按多单、市价加 1% 平仓来估的。外汇和贵金属杠杆高，这种默认上浮若碰上跳空可能低估真实回撤，上 MT5 自己改 ClosePrice 逻辑更稳。 手数落地靠 Volume() 先按风险配置取粗值，再经 ValidateLotsize() 和 NormalizeLotsize() 两道关。前者按 RiskFloor 切：RiskFloorMin 低于最小手数直接钳到最小；RiskFloorMax 超了最大可承险就退到「剩余手数的 RiskFloorPercentage 与最小手数取大」。后者查品种交易量限制和步长，违规则算剩余可用手数再走一遍下限判断。 GetRisk() 的算法很直接：先由 GetMinimumRisk() 算出最小手数对应的最小风险金额 MinimumAmount，再用指定 Amount 除以它得商，乘最小手数得目标手数。这种除法映射在迷你账户里若 MinimumAmount 极小，容易出现手数跳变，建议打印 MinimumAmount 看数量级。
class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| NewsTrading | class=class="str">"cmt">//| Copyright class="num">2024, MetaQuotes Ltd. | class=class="str">"cmt">//| [MQL5官方文档] | class=class="str">"cmt">//+------------------------------------------------------------------+ class="macro">#include "ChartProperties.mqh" class="macro">#include <Trade/AccountInfo.mqh> CAccountInfo Account; class=class="str">"cmt">//-- Enumeration declaration for Risk options enum RiskOptions { MINIMUM_LOT,class=class="str">"cmt">//MINIMUM LOTSIZE MAXIMUM_LOT,class=class="str">"cmt">//MAXIMUM LOTSIZE PERCENTAGE_OF_BALANCE,class=class="str">"cmt">//PERCENTAGE OF BALANCE PERCENTAGE_OF_FREEMARGIN,class=class="str">"cmt">//PERCENTAGE OF FREE-MARGIN AMOUNT_PER_BALANCE,class=class="str">"cmt">//AMOUNT PER BALANCE AMOUNT_PER_FREEMARGIN,class=class="str">"cmt">//AMOUNT PER FREE-MARGIN LOTSIZE_PER_BALANCE,class=class="str">"cmt">//LOTSIZE PER BALANCE LOTSIZE_PER_FREEMARGIN,class=class="str">"cmt">//LOTSIZE PER FREE-MARGIN CUSTOM_LOT,class=class="str">"cmt">//CUSTOM LOTSIZE PERCENTAGE_OF_MAXRISKclass=class="str">"cmt">//PERCENTAGE OF MAX-RISK } RiskProfileOption;class=class="str">"cmt">//variable for Risk options class=class="str">"cmt">//-- Enumeration declaration for Risk floor enum RiskFloor { RiskFloorMin,class=class="str">"cmt">//MINIMUM LOTSIZE RiskFloorMax,class=class="str">"cmt">//MAX-RISK RiskFloorNoneclass=class="str">"cmt">//NONE } RiskFloorOption;class=class="str">"cmt">//variable for Risk floor class=class="str">"cmt">//-- Enumeration declaration for Risk ceiling(Maximum allowable risk in terms of lot-size) enum RiskCeil { RiskCeilMax,class=class="str">"cmt">//MAX LOTSIZE RiskCeilMax2,class=class="str">"cmt">//MAX LOTSIZE(x2) RiskCeilMax3,class=class="str">"cmt">//MAX LOTSIZE(x3) RiskCeilMax4,class=class="str">"cmt">//MAX LOTSIZE(x4)
◍ 风控模块的账户基准与仓位结构
在 MT5 的 EA 风控类里，账户基准通常抽象成 BALANCE 与 MARGIN 两种介质。下面这段声明把「按余额/空闲保证金计算」的多种风险剖面拆成了独立 struct，方便运行时切换而不必重写主逻辑。 RiskCeilOption 枚举里放了 RiskCeilMax5 作为手数上限倍率（x5），实际含义是：当风控触发上限时，允许开仓手数不超过基准的 5 倍。RISK_AMOUNT 与 RISK_LOT 两个 struct 分别存「金额」和「手数」剖面，各自带一个 RiskXXXBoF 字段标记当前用的是 Balance 还是 Free-Margin。 CRiskManagement 类私有区有 max_percent 常量，用于锁死最大风险百分比；RiskProfile1~3 三个方法对应三种计算模式，入参都是 RiskMedium 枚举。ORDERTYPE 用 ENUM_ORDER_TYPE 存方向，配合 OpenPrice / ClosePrice 做后续盈亏比换算。 开 MT5 新建类时，建议先照这个结构把 RiskMedium 枚举和三个 RiskProfile 方法骨架敲出来，再填计算体，能少走很多弯路。外汇与贵金属杠杆高，这类仓位计算任何偏差都可能放大亏损。
RiskCeilMax5,class=class="str">"cmt">//MAX LOTSIZE(x5) } RiskCeilOption;class=class="str">"cmt">//variable for Risk ceiling class=class="str">"cmt">//-- Structure declaration for Risk options(AMOUNT PER BALANCE and AMOUNT PER FREE-MARGIN) class="kw">struct RISK_AMOUNT { class="type">class="kw">double RiskAmountBoF;class=class="str">"cmt">//store Balance or Free-Margin class="type">class="kw">double RiskAmount;class=class="str">"cmt">//store risk amount } Risk_Profile_2;class=class="str">"cmt">//variable for Risk options(AMOUNT PER BALANCE and AMOUNT PER FREE-MARGIN) class=class="str">"cmt">//-- Structure declaration for Risk options(LOTSIZE PER BALANCE and LOTSIZE PER FREE-MARGIN) class="kw">struct RISK_LOT { class="type">class="kw">double RiskLotBoF;class=class="str">"cmt">//store Balance or Free-Margin class="type">class="kw">double RiskLot;class=class="str">"cmt">//store lot-size } Risk_Profile_3;class=class="str">"cmt">//variable for Risk options(LOTSIZE PER BALANCE and LOTSIZE PER FREE-MARGIN) class="type">class="kw">double RiskFloorPercentage;class=class="str">"cmt">//variable for RiskFloorMax class="type">class="kw">double Risk_Profile_1;class=class="str">"cmt">//variable for Risk options(PERCENTAGE OF BALANCE and PERCENTAGE OF FREE-MARGIN) class="type">class="kw">double Risk_Profile_4;class=class="str">"cmt">//variable for Risk option(CUSTOM LOTSIZE) class="type">class="kw">double Risk_Profile_5;class=class="str">"cmt">//variable for Risk option(PERCENTAGE OF MAX-RISK) class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|RiskManagement class | class=class="str">"cmt">//+------------------------------------------------------------------+ class CRiskManagement : class="kw">public CChartProperties { class="kw">private: class="type">class="kw">double Medium;class=class="str">"cmt">//variable to store actual Account(Balance or Free-Margin) class="type">class="kw">double RiskAmount,MinimumAmount; class="type">class="kw">double Lots;class=class="str">"cmt">//variable to store Lot-size to open trade const class="type">class="kw">double max_percent;class=class="str">"cmt">//variable to store percentage for Maximum risk class=class="str">"cmt">//-- enumeration for dealing with account balance/free-margin enum RiskMedium { BALANCE, MARGIN }; class=class="str">"cmt">//-- calculations for Risk options(PERCENTAGE OF BALANCE and PERCENTAGE OF FREE-MARGIN) class="type">class="kw">double RiskProfile1(const RiskMedium R_Medium); class=class="str">"cmt">//-- calculations for Risk options(AMOUNT PER BALANCE and AMOUNT PER FREE-MARGIN) class="type">class="kw">double RiskProfile2(const RiskMedium R_Medium); class=class="str">"cmt">//-- calculations for Risk options(LOTSIZE PER BALANCE and LOTSIZE PER FREE-MARGIN) class="type">class="kw">double RiskProfile3(const RiskMedium R_Medium); class=class="str">"cmt">//-- calculations for Maximum allowable Risk class="type">class="kw">double MaxRisk(const class="type">class="kw">double percent); class=class="str">"cmt">//-- Store Trade&class="macro">#x27;s Open-price class="type">class="kw">double OpenPrice; class=class="str">"cmt">//-- Store Trade&class="macro">#x27;s Close-price class="type">class="kw">double ClosePrice; class=class="str">"cmt">//-- Store Ordertype between(ORDER_TYPE_BUY or ORDER_TYPE_SELL) for risk calaculations ENUM_ORDER_TYPE ORDERTYPE; class=class="str">"cmt">//-- Set Medium variable value
风控类里的手数推导与订单类型判定
这段 C++ 风格封装把仓位计算的底层逻辑收进了 CRiskManagement 类。先用 SetMedium 决定风险基数取自账户余额还是可用保证金：传入 BALANCE 就抓 Account.Balance()，否则取 Account.FreeMargin()，这一步直接决定后面百分比算法的分母。 GetMinimumRisk 借助 OrderCalcProfit，以经纪商返回的最小手数 LotsMin() 试算一笔平仓盈亏，结果写进 MinimumAmount。若最小手数都算不出利润或传入金额为 0，GetRisk 直接返回 0.0；否则按 (Amount/MinimumAmount)*LotsMin() 线性外推出对应手数，属于最朴素的比例缩放。 SetOrderType 把市价单和挂单统一归并：BUY / BUY_LIMIT / BUY_STOP 全映射为 ORDER_TYPE_BUY，SELL 系列同理。这样后面算保证金和利润时只需判断多空两个分支，不用每种订单写一套。 GetRiskOption 用 switch 把枚举转成可读字符串，覆盖了 MINIMUM_LOT、MAXIMUM_LOT、PERCENTAGE_OF_BALANCE、PERCENTAGE_OF_FREEMARGIN、AMOUNT_PER_BALANCE 五种配置。实盘里外汇和贵金属杠杆高、滑点跳空频繁，这类手数推导只是仓位上限参考，真实成交可能偏离，开 MT5 用策略测试器跑一下 LotsMin() 返回值最踏实。
class="type">void SetMedium(const RiskMedium R_Medium) {Medium = (R_Medium==BALANCE)?Account.Balance():Account.FreeMargin();} class=class="str">"cmt">//-- Get Minimum Risk for a Trade using Minimum Lot-size class="type">bool GetMinimumRisk() { class="kw">return OrderCalcProfit(ORDERTYPE,Symbol(),LotsMin(),OpenPrice,ClosePrice,MinimumAmount); } class=class="str">"cmt">//-- Retrieve Risk amount based on Risk inputs class="type">class="kw">double GetRisk(class="type">class="kw">double Amount) { if(!GetMinimumRisk()||Amount==class="num">0) class="kw">return class="num">0.0; class="kw">return ((Amount/MinimumAmount)*LotsMin()); } class="kw">protected: class=class="str">"cmt">//-- Application of Lot-size limits class="type">void ValidateLotsize(class="type">class="kw">double &Lotsize); class=class="str">"cmt">//-- Set ORDERTYPE variable to(ORDER_TYPE_BUY or ORDER_TYPE_SELL) respectively class="type">void SetOrderType(ENUM_ORDER_TYPE Type) { if(Type==ORDER_TYPE_BUY||Type==ORDER_TYPE_BUY_LIMIT||Type==ORDER_TYPE_BUY_STOP) { ORDERTYPE = ORDER_TYPE_BUY; } else if(Type==ORDER_TYPE_SELL||Type==ORDER_TYPE_SELL_LIMIT||Type==ORDER_TYPE_SELL_STOP) { ORDERTYPE = ORDER_TYPE_SELL; } } class="kw">public: CRiskManagement();class=class="str">"cmt">//Class&class="macro">#x27;s constructor class=class="str">"cmt">//-- Retrieve user&class="macro">#x27;s Risk option class="type">class="kw">string GetRiskOption() { class="kw">switch(RiskProfileOption) { case MINIMUM_LOT:class=class="str">"cmt">//MINIMUM LOTSIZE - Risk Option class="kw">return "MINIMUM LOTSIZE"; class="kw">break; case MAXIMUM_LOT:class=class="str">"cmt">//MAXIMUM LOTSIZE - Risk Option class="kw">return "MAXIMUM LOTSIZE"; class="kw">break; case PERCENTAGE_OF_BALANCE:class=class="str">"cmt">//PERCENTAGE OF BALANCE - Risk Option class="kw">return "PERCENTAGE OF BALANCE"; class="kw">break; case PERCENTAGE_OF_FREEMARGIN:class=class="str">"cmt">//PERCENTAGE OF FREE-MARGIN - Risk Option class="kw">return "PERCENTAGE OF FREE-MARGIN"; class="kw">break; case AMOUNT_PER_BALANCE:class=class="str">"cmt">//AMOUNT PER BALANCE - Risk Option
「仓位风险档位的字符串映射」
这段 MQL5 片段把枚举型的风险配置选项转成可读字符串，方便在面板或日志里直接输出当前策略用的风控档位。 第一个 switch 处理主风险模式：AMOUNT_PER_BALANCE 对应按账户余额算金额，AMOUNT_PER_FREEMARGIN 按空闲保证金算，LOTSIZE_PER_BALANCE 与 LOTSIZE_PER_FREEMARGIN 则改为按手数比例，CUSTOM_LOT 是写死手数，PERCENTAGE_OF_MAXRISK 按最大可承受风险的百分比下单。 下方 GetRiskFloor() 管风险下限：RiskFloorMin 输出 MINIMUM LOTSIZE，RiskFloorMax 输出 MAX-RISK，RiskFloorNone 就是 NONE，没匹配到返回空串。GetRiskCeil() 管上限，RiskCeilMax 返回 MAX LOTSIZE，RiskCeilMax2 返回 MAX LOTSIZE(x2) 表示允许双倍上限。 在 MT5 里把这段贴进 EA 的调试函数，能立刻看到你当前加载的仓位风控到底是按余额还是按保证金在跑；外汇和贵金属杠杆高，映射错档位可能让实际敞口偏离预期。
class="kw">return "AMOUNT PER BALANCE"; class="kw">break; case AMOUNT_PER_FREEMARGIN:class=class="str">"cmt">//AMOUNT PER FREE-MARGIN - Risk Option class="kw">return "AMOUNT PER FREE-MARGIN"; class="kw">break; case LOTSIZE_PER_BALANCE:class=class="str">"cmt">//LOTSIZE PER BALANCE - Risk Option class="kw">return "LOTSIZE PER BALANCE"; class="kw">break; case LOTSIZE_PER_FREEMARGIN:class=class="str">"cmt">//LOTSIZE PER FREE-MARGIN - Risk Option class="kw">return "LOTSIZE PER FREE-MARGIN"; class="kw">break; case CUSTOM_LOT:class=class="str">"cmt">//CUSTOM LOTSIZE - Risk Option class="kw">return "CUSTOM LOTSIZE"; class="kw">break; case PERCENTAGE_OF_MAXRISK:class=class="str">"cmt">//PERCENTAGE OF MAX-RISK - Risk Option class="kw">return "PERCENTAGE OF MAX-RISK"; class="kw">break; class="kw">default: class="kw">return ""; class="kw">break; } } class=class="str">"cmt">//-- Retrieve user&class="macro">#x27;s Risk Floor Option class="type">class="kw">string GetRiskFloor() { class="kw">switch(RiskFloorOption) { case RiskFloorMin:class=class="str">"cmt">//MINIMUM LOTSIZE for Risk floor options class="kw">return "MINIMUM LOTSIZE"; class="kw">break; case RiskFloorMax:class=class="str">"cmt">//MAX-RISK for Risk floor options class="kw">return "MAX-RISK"; class="kw">break; case RiskFloorNone:class=class="str">"cmt">//NONE for Risk floor options class="kw">return "NONE"; class="kw">break; class="kw">default: class="kw">return ""; class="kw">break; } } class=class="str">"cmt">//-- Retrieve user&class="macro">#x27;s Risk Ceiling option class="type">class="kw">string GetRiskCeil() { class="kw">switch(RiskCeilOption) { case RiskCeilMax:class=class="str">"cmt">//MAX LOTSIZE for Risk ceiling options class="kw">return "MAX LOTSIZE"; class="kw">break; case RiskCeilMax2:class=class="str">"cmt">//MAX LOTSIZE(x2) for Risk ceiling options class="kw">return "MAX LOTSIZE(x2)";
◍ 风险档位如何映射成手数返回值
这段 CRiskManagement 类的枚举分支，把风控档位直接翻译成 MT5 可读的手数字符串或数值。RiskCeilMax3 到 RiskCeilMax5 三个 case 分别返回 "MAX LOTSIZE(x3)"、"MAX LOTSIZE(x4)"、"MAX LOTSIZE(x5)"，对应风险上限选项中的 3 倍、4 倍、5 倍最大仓位倍数；default 返回空串，避免越界调用崩在运行时。 Volume() 才是真正算手数的入口。MINIMUM_LOT 直接取 LotsMin()，MAXIMUM_LOT 取 LotsMax()；PERCENTAGE_OF_BALANCE 与 PERCENTAGE_OF_FREEMARGIN 走 RiskProfile1()，区别只是传 BALANCE 还是 MARGIN；AMOUNT_PER_* 走 RiskProfile2()，LOTSIZE_PER_BALANCE 走 RiskProfile3(BALANCE)。 构造函数里 OpenPrice 初始化为 Ask()，ClosePrice 用 NormalizePrice(Ask()+Ask()*0.01) 预设了向上 1% 的平仓参考价。你可以在 MT5 里新建 EA 继承这个类，把 RiskProfileOption 改成 LOTSIZE_PER_BALANCE，看回测里手数是否随余额线性变化——外汇和贵金属杠杆高，实测前先关掉自动下单。
class="kw">break; case RiskCeilMax3:class=class="str">"cmt">//MAX LOTSIZE(x3) for Risk ceiling options class="kw">return "MAX LOTSIZE(x3)"; class="kw">break; case RiskCeilMax4:class=class="str">"cmt">//MAX LOTSIZE(x4) for Risk ceiling options class="kw">return "MAX LOTSIZE(x4)"; class="kw">break; case RiskCeilMax5:class=class="str">"cmt">//MAX LOTSIZE(x5) for Risk ceiling options class="kw">return "MAX LOTSIZE(x5)"; class="kw">break; class="kw">default: class="kw">return ""; class="kw">break; } } class="type">class="kw">double Volume();class=class="str">"cmt">//Get risk in Volume class=class="str">"cmt">//Apply fixes to lot-size where applicable class="type">void NormalizeLotsize(class="type">class="kw">double &Lotsize); }; class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Constructor | class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//Initialize values CRiskManagement::CRiskManagement(class="type">void):Lots(class="num">0.0),max_percent(class="num">100), ORDERTYPE(ORDER_TYPE_BUY),OpenPrice(Ask()), ClosePrice(NormalizePrice(Ask()+Ask()*class="num">0.01)) { } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|Get risk in Volume | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">class="kw">double CRiskManagement::Volume() { class="kw">switch(RiskProfileOption) { case MINIMUM_LOT:class=class="str">"cmt">//MINIMUM LOTSIZE - Risk Option class="kw">return LotsMin(); class="kw">break; case MAXIMUM_LOT:class=class="str">"cmt">//MAXIMUM LOTSIZE - Risk Option Lots = LotsMax(); class="kw">break; case PERCENTAGE_OF_BALANCE:class=class="str">"cmt">//PERCENTAGE OF BALANCE - Risk Option Lots = RiskProfile1(BALANCE); class="kw">break; case PERCENTAGE_OF_FREEMARGIN:class=class="str">"cmt">//PERCENTAGE OF FREE-MARGIN - Risk Option Lots = RiskProfile1(MARGIN); class="kw">break; case AMOUNT_PER_BALANCE:class=class="str">"cmt">//AMOUNT PER BALANCE - Risk Option Lots = RiskProfile2(BALANCE); class="kw">break; case AMOUNT_PER_FREEMARGIN:class=class="str">"cmt">//AMOUNT PER FREE-MARGIN - Risk Option Lots = RiskProfile2(MARGIN); class="kw">break; case LOTSIZE_PER_BALANCE:class=class="str">"cmt">//LOTSIZE PER BALANCE - Risk Option Lots = RiskProfile3(BALANCE);
仓位计算的分支出口与风险函数实现
上面这段 switch 结构把五种风险模式映射到具体的手数计算函数，最后统一走 ValidateLotsize 与 NormalizeLotsize 做边界修正和小数点规范化，再 return 出手数。 PERCENTAGE_OF_BALANCE 与 PERCENTAGE_OF_FREEMARGIN 共用 RiskProfile1：先把基准 Medium（余额或空闲保证金）乘以 Risk_Profile_1 的百分数得出 RiskAmount，再交给 GetRisk 换算手数。 RiskProfile2 处理固定金额类模式，用 RiskAmountBoF 与 RiskAmount 的比值做缩放，且强制比值不低于 1；若 Medium≤0 直接返回 0.0，避免空账户开仓。 RiskProfile3 针对「每单位保证金对应手数」的逻辑，Medium>0 时按 RiskLotBoF 与 RiskLot 比例放大，否则返回 0.0。 MaxRisk 是最大可承受风险入口，入参 percent 超出 [0.01,100] 会 Print 报错并返回 0.0；内部调用 OrderCalcMargin(ORDERTYPE,Symbol(),1.0,OpenPrice,margin) 先算 1 手保证金占用，作为后续风控基准。外汇与贵金属杠杆波动大，这类计算仅降低爆仓概率，不消除高风险。
class="kw">break; case LOTSIZE_PER_FREEMARGIN:class=class="str">"cmt">//LOTSIZE PER FREE-MARGIN - Risk Option Lots = RiskProfile3(MARGIN); class="kw">break; case CUSTOM_LOT:class=class="str">"cmt">//CUSTOM LOTSIZE - Risk Option Lots = Risk_Profile_4; class="kw">break; case PERCENTAGE_OF_MAXRISK:class=class="str">"cmt">//PERCENTAGE OF MAX-RISK - Risk Option Lots = MaxRisk(Risk_Profile_5); class="kw">break; class="kw">default: Lots = class="num">0.0; class="kw">break; } ValidateLotsize(Lots);class=class="str">"cmt">//Check/Adjust Lotsize Limits NormalizeLotsize(Lots);class=class="str">"cmt">//Normalize Lotsize class="kw">return Lots; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|calculations for Risk options | class=class="str">"cmt">//|(PERCENTAGE OF BALANCE and PERCENTAGE OF FREE-MARGIN) | class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//-- calculations for Risk options(PERCENTAGE OF BALANCE and PERCENTAGE OF FREE-MARGIN) class="type">class="kw">double CRiskManagement::RiskProfile1(const RiskMedium R_Medium) { SetMedium(R_Medium); RiskAmount = Medium*(Risk_Profile_1/class="num">100); class="kw">return GetRisk(RiskAmount); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|calculations for Risk options | class=class="str">"cmt">//|(AMOUNT PER BALANCE and AMOUNT PER FREE-MARGIN) | class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//-- calculations for Risk options(AMOUNT PER BALANCE and AMOUNT PER FREE-MARGIN) class="type">class="kw">double CRiskManagement::RiskProfile2(const RiskMedium R_Medium) { SetMedium(R_Medium); class="type">class="kw">double risk = (Risk_Profile_2.RiskAmountBoF/Risk_Profile_2.RiskAmount); risk = (risk<class="num">1)?class="num">1:risk; if(Medium<=class="num">0) class="kw">return class="num">0.0; RiskAmount = Medium/risk; class="kw">return GetRisk(RiskAmount); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|calculations for Risk options | class=class="str">"cmt">//|(LOTSIZE PER BALANCE and LOTSIZE PER FREE-MARGIN) | class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//-- calculations for Risk options(LOTSIZE PER BALANCE and LOTSIZE PER FREE-MARGIN) class="type">class="kw">double CRiskManagement::RiskProfile3(const RiskMedium R_Medium) { SetMedium(R_Medium); class="kw">return (Medium>class="num">0)?((Medium/Risk_Profile_3.RiskLotBoF)*Risk_Profile_3.RiskLot):class="num">0.0; } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//|calculations for Maximum allowable Risk | class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//-- calculations for Maximum allowable Risk class="type">class="kw">double CRiskManagement::MaxRisk(const class="type">class="kw">double percent) { class="type">class="kw">double margin=class="num">0.0,max_risk=class="num">0.0; class=class="str">"cmt">//--- checks if(percent<class="num">0.01 || percent>class="num">100) { Print(__FUNCTION__," invalid parameters"); class="kw">return(class="num">0.0); } class=class="str">"cmt">//--- calculate margin requirements for class="num">1 lot if(!OrderCalcMargin(ORDERTYPE,Symbol(),class="num">1.0,OpenPrice,margin) || margin<class="num">0.0) {
「手数封顶与步进对齐的兜底逻辑」
风险模块算出手数后，不能直接拿去下单。交易所或经纪商对单品种有总持仓上限（LotsLimit），也规定了最小手数（LotsMin）和步进（LotsStep），忽略这些字段会被 MT5 拒单。 当已开仓位加挂单量逼近 LotsLimit 时，代码先算 remaining_avail_lots = LotsLimit - (PositionsVolume + OrdersVolume)。若剩余量大于等于 LotsMin，按风险下限模式二选一：选 RiskFloorMin 就硬切到 LotsMin 并打警告；选 RiskFloorMax 则按剩余量的 RiskFloorPercentage% 缩手数，且不低于 LotsMin。 若剩余量连 LotsMin 都够不上，Lotsize 直接置 0 并返回，等价于这轮不下单。最后一段用 LotsStep * MathFloor(Lotsize / LotsStep) 把任意手数吸附到合规步进格点上——这一步不写，0.13 手这类值在高点差贵金属上大概率报无效交易量。 外汇与贵金属杠杆波动剧烈，总限仓逻辑只是防拒单，不表示风险可控，实盘前请在策略测试器用真实点差跑一遍 NormalizeLotsize。
class="type">void CRiskManagement::NormalizeLotsize(class="type">class="kw">double &Lotsize) { if(Lotsize<=class="num">0.0) class="kw">return; class=class="str">"cmt">//-- Check if the is a Volume limit for the current Symbol if(LotsLimit()>class="num">0.0) { if((Lots+PositionsVolume()+OrdersVolume())>LotsLimit()) { class=class="str">"cmt">//-- calculation of available lotsize remaining class="type">class="kw">double remaining_avail_lots = (LotsLimit()-(PositionsVolume()+OrdersVolume())); if(remaining_avail_lots>=LotsMin()) { if(RiskFloorOption==RiskFloorMin)class=class="str">"cmt">//Check if Risk floor option is MINIMUM LOTSIZE { Print("Warning: Volume Limit Reached, minimum Lotsize selected."); Lotsize = LotsMin(); } else if(RiskFloorOption==RiskFloorMax)class=class="str">"cmt">//Check if Risk floor option is MAX-RISK { Print("Warning: Volume Limit Reached, Lotsize Reduced."); Lotsize = ((remaining_avail_lots*(RiskFloorPercentage/class="num">100))>LotsMin())? (remaining_avail_lots*(RiskFloorPercentage/class="num">100)):LotsMin(); } } else { Print("Volume Limit Reached!"); Lotsize=class="num">0.0; class="kw">return; } } } class=class="str">"cmt">//Check if there is a valid Volume Step for the current Symbol if(LotsStep()>class="num">0.0) Lotsize=LotsStep()*MathFloor(Lotsize/LotsStep()); }
◍ 手数下限与上限的硬约束逻辑
风控模块里手数不是算出来就直接发单，得先过两道闸： floor 管下限、ceil 管上限。floor 有三种模式——Min 模式直接把低于 broker 最小手数或超最大风险的手数压回 LotsMin()；Max 模式允许手数顶到 RiskFloorPercentage 算出的风险上限，但绝不会低于最小手数；None 模式则狠一点，低于最小手数直接归零不发单。 上限侧同样分档：RiskCeilMax 卡在 LotsMax()，Max2、Max3 分别放宽到 2 倍、3 倍券商最大手数。实盘里黄金 XAUUSD 的 LotsMax() 常见为 50，若策略算出 120 手且开了 Max3，会被截断到 150 以内但超出 50 的单笔暴露仍偏高，外汇与贵金属杠杆风险极大，参数需按账户净值倒推。 下面这段是 MT5 里 ValidateLotsize 的核心分支，建议直接拷进 EA 的 CRiskManagement 类里跑一遍。
class="type">void CRiskManagement::ValidateLotsize(class="type">class="kw">double &Lotsize) { class="kw">switch(RiskFloorOption) { case RiskFloorMin:class=class="str">"cmt">//MINIMUM LOTSIZE for Risk floor options class=class="str">"cmt">//-- Check if lot-size is not less than Minimum lot or more than maximum allowable risk if(Lotsize<LotsMin()||Lotsize>MaxRisk(max_percent)) { Lotsize=LotsMin(); } class="kw">break; case RiskFloorMax:class=class="str">"cmt">//MAX-RISK for Risk floor options class=class="str">"cmt">//-- Check if lot-size is more the maximum allowable risk if(Lotsize>MaxRisk(max_percent)) { Lotsize=(MaxRisk(RiskFloorPercentage)>LotsMin())?MaxRisk(RiskFloorPercentage):LotsMin(); } else if(Lotsize<LotsMin())class=class="str">"cmt">//Check if lot-size is less than Minimum lot { Lotsize=LotsMin(); } class="kw">break; case RiskFloorNone:class=class="str">"cmt">//NONE for Risk floor options class=class="str">"cmt">//Check if lot-size is less than Minimum lot if(Lotsize<LotsMin()) { Lotsize=class="num">0.0; } class="kw">break; class="kw">default: Lotsize=class="num">0.0; class="kw">break; } class="kw">switch(RiskCeilOption) { case RiskCeilMax:class=class="str">"cmt">//MAX LOTSIZE for Risk ceiling options class=class="str">"cmt">//Check if lot-size is more than Maximum lot if(Lotsize>LotsMax()) Lotsize=LotsMax(); class="kw">break; case RiskCeilMax2:class=class="str">"cmt">//MAX LOTSIZE(x2) for Risk ceiling options class=class="str">"cmt">//Check if lot-size is more than Maximum lot times two if(Lotsize>(LotsMax()*class="num">2)) Lotsize=(LotsMax()*class="num">2); class="kw">break; case RiskCeilMax3:class=class="str">"cmt">//MAX LOTSIZE(x3) for Risk ceiling options class=class="str">"cmt">//Check if lot-size is more than Maximum lot times three if(Lotsize>(LotsMax()*class="num">3))
手数封顶的多档熔断逻辑
在 EA 风控里，RiskCeil 枚举把「手数上限」拆成了 5 档：从经纪商最大允许手数 LotsMax() 的 1 倍到 5 倍。switch 里每个 case 只做一件事——若计算出的 Lotsize 超过该档倍数，就截断成 LotsMax() 乘对应系数，然后 break。 比如 RiskCeilMax3 档，超过 3 倍最大手数就强制压回 3 倍；RiskCeilMax5 则放到 5 倍才截。这种写法让仓位上限可配置，而不是写死一个常数。 配套的 RiskOptions 枚举列了 10 种风险计算源：最小手数、最大手数、余额百分比、可用保证金百分比、每单位余额金额等，说明仓位不是拍脑袋来的。外汇和贵金属杠杆高，这类封顶逻辑能避免极端波动下保证金瞬间击穿，但市场仍可能跳空导致实际风险超预期。 开 MT5 把下面这段代码塞进你的仓位模块，改 RiskCeilOption 的值就能切换封顶倍数，立刻能看到手数被钳在哪一档。
case RiskCeilMax3:class=class="str">"cmt">//MAX LOTSIZE(x3) for Risk ceiling options class=class="str">"cmt">//Check if lot-size is more than Maximum lot times three if(Lotsize>(LotsMax()*class="num">3)) Lotsize=(LotsMax()*class="num">3); class="kw">break; case RiskCeilMax4:class=class="str">"cmt">//MAX LOTSIZE(x4) for Risk ceiling options class=class="str">"cmt">//Check if lot-size is more than Maximum lot times four if(Lotsize>(LotsMax()*class="num">4)) Lotsize=(LotsMax()*class="num">4); class="kw">break; case RiskCeilMax5:class=class="str">"cmt">//MAX LOTSIZE(x5) for Risk ceiling options class=class="str">"cmt">//Check if lot-size is more than Maximum lot times five if(Lotsize>(LotsMax()*class="num">5)) Lotsize=(LotsMax()*class="num">5); class="kw">break; class="kw">default: class="kw">break; } } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//-- Enumeration declaration for Risk options enum RiskOptions { MINIMUM_LOT,class=class="str">"cmt">//MINIMUM LOTSIZE MAXIMUM_LOT,class=class="str">"cmt">//MAXIMUM LOTSIZE PERCENTAGE_OF_BALANCE,class=class="str">"cmt">//PERCENTAGE OF BALANCE PERCENTAGE_OF_FREEMARGIN,class=class="str">"cmt">//PERCENTAGE OF FREE-MARGIN AMOUNT_PER_BALANCE,class=class="str">"cmt">//AMOUNT PER BALANCE AMOUNT_PER_FREEMARGIN,class=class="str">"cmt">//AMOUNT PER FREE-MARGIN LOTSIZE_PER_BALANCE,class=class="str">"cmt">//LOTSIZE PER BALANCE LOTSIZE_PER_FREEMARGIN,class=class="str">"cmt">//LOTSIZE PER FREE-MARGIN CUSTOM_LOT,class=class="str">"cmt">//CUSTOM LOTSIZE PERCENTAGE_OF_MAXRISKclass=class="str">"cmt">//PERCENTAGE OF MAX-RISK } RiskProfileOption;class=class="str">"cmt">//variable for Risk options class=class="str">"cmt">//-- Enumeration declaration for Risk floor enum RiskFloor { RiskFloorMin,class=class="str">"cmt">//MINIMUM LOTSIZE RiskFloorMax,class=class="str">"cmt">//MAX-RISK RiskFloorNoneclass=class="str">"cmt">//NONE } RiskFloorOption;class=class="str">"cmt">//variable for Risk floor class=class="str">"cmt">//-- Enumeration declaration for Risk ceiling(Maximum allowable risk in terms of lot-size) enum RiskCeil { RiskCeilMax,class=class="str">"cmt">//MAX LOTSIZE RiskCeilMax2,class=class="str">"cmt">//MAX LOTSIZE(x2) RiskCeilMax3,class=class="str">"cmt">//MAX LOTSIZE(x3) RiskCeilMax4,class=class="str">"cmt">//MAX LOTSIZE(x4) RiskCeilMax5,class=class="str">"cmt">//MAX LOTSIZE(x5) } RiskCeilOption;class=class="str">"cmt">//variable for Risk ceiling class=class="str">"cmt">//-- Structure declaration for Risk options(AMOUNT PER BALANCE and AMOUNT PER FREE-MARGIN) class="kw">struct RISK_AMOUNT { class="type">class="kw">double RiskAmountBoF;class=class="str">"cmt">//store Balance or Free-Margin class="type">class="kw">double RiskAmount;class=class="str">"cmt">//store risk amount } Risk_Profile_2;class=class="str">"cmt">//variable for Risk options(AMOUNT PER BALANCE and AMOUNT PER FREE-MARGIN) class=class="str">"cmt">//-- Structure declaration for Risk options(LOTSIZE PER BALANCE and LOTSIZE PER FREE-MARGIN) class="kw">struct RISK_LOT { class="type">class="kw">double RiskLotBoF;class=class="str">"cmt">//store Balance or Free-Margin class="type">class="kw">double RiskLot;class=class="str">"cmt">//store lot-size } Risk_Profile_3;class=class="str">"cmt">//variable for Risk options(LOTSIZE PER BALANCE and LOTSIZE PER FREE-MARGIN)
「把风险档位映射成可读字符串」
EA 里风险档位通常用枚举值驱动，但日志和面板只认人话。下面这段代码把 RiskProfileOption 的枚举分支逐一转成字符串，方便在测试报告里直接看出当前跑的是哪一档。 double RiskFloorPercentage; // 风险下限占余额的百分比变量 double Risk_Profile_1; // 对应「余额百分比」与「空闲保证金百分比」两档风险的变量 double Risk_Profile_4; // 对应「自定义手数」档的变量 double Risk_Profile_5; // 对应「最大风险百分比」档的变量 GetRiskOption() 用 switch 把 10 种枚举都覆盖了：从 MINIMUM_LOT、MAXIMUM_LOT 到 PERCENTAGE_OF_BALANCE、PERCENTAGE_OF_FREEMARGIN，再到 AMOUNT_PER_BALANCE、LOTSIZE_PER_FREEMARGIN 以及 CUSTOM_LOT、PERCENTAGE_OF_MAXRISK，default 返回空串。 实盘接 MT5 时，把这个函数挂到 OnTester 或面板刷新里，回测 100 次就能在 CSV 里看到每单对应的风险模式，不用反查枚举表。外汇与贵金属杠杆高，同一套 RiskProfileOption 在 1:500 和 1:30 下爆仓概率可能差数倍，参数切换前先跑一遍历史样本。
class="type">class="kw">double RiskFloorPercentage; class=class="str">"cmt">//variable for RiskFloorMax class="type">class="kw">double Risk_Profile_1; class=class="str">"cmt">//variable for Risk options(PERCENTAGE OF BALANCE and PERCENTAGE OF FREE-MARGIN) class="type">class="kw">double Risk_Profile_4; class=class="str">"cmt">//variable for Risk option(CUSTOM LOTSIZE) class="type">class="kw">double Risk_Profile_5; class=class="str">"cmt">//variable for Risk option(PERCENTAGE OF MAX-RISK) class=class="str">"cmt">//-- Retrieve user&class="macro">#x27;s Risk option class="type">class="kw">string GetRiskOption() { class="kw">switch(RiskProfileOption) { case MINIMUM_LOT: class=class="str">"cmt">//MINIMUM LOTSIZE - Risk Option class="kw">return "MINIMUM LOTSIZE"; class="kw">break; case MAXIMUM_LOT: class=class="str">"cmt">//MAXIMUM LOTSIZE - Risk Option class="kw">return "MAXIMUM LOTSIZE"; class="kw">break; case PERCENTAGE_OF_BALANCE: class=class="str">"cmt">//PERCENTAGE OF BALANCE - Risk Option class="kw">return "PERCENTAGE OF BALANCE"; class="kw">break; case PERCENTAGE_OF_FREEMARGIN:class=class="str">"cmt">//PERCENTAGE OF FREE-MARGIN - Risk Option class="kw">return "PERCENTAGE OF FREE-MARGIN"; class="kw">break; case AMOUNT_PER_BALANCE: class=class="str">"cmt">//AMOUNT PER BALANCE - Risk Option class="kw">return "AMOUNT PER BALANCE"; class="kw">break; case AMOUNT_PER_FREEMARGIN: class=class="str">"cmt">//AMOUNT PER FREE-MARGIN - Risk Option class="kw">return "AMOUNT PER FREE-MARGIN"; class="kw">break; case LOTSIZE_PER_BALANCE: class=class="str">"cmt">//LOTSIZE PER BALANCE - Risk Option class="kw">return "LOTSIZE PER BALANCE"; class="kw">break; case LOTSIZE_PER_FREEMARGIN:class=class="str">"cmt">//LOTSIZE PER FREE-MARGIN - Risk Option class="kw">return "LOTSIZE PER FREE-MARGIN"; class="kw">break; case CUSTOM_LOT: class=class="str">"cmt">//CUSTOM LOTSIZE - Risk Option class="kw">return "CUSTOM LOTSIZE"; class="kw">break; case PERCENTAGE_OF_MAXRISK: class=class="str">"cmt">//PERCENTAGE OF MAX-RISK - Risk Option class="kw">return "PERCENTAGE OF MAX-RISK"; class="kw">break; class="kw">default: class="kw">return "";
◍ 风险上下限的字符串映射与初始化
仓位风控模块里，Risk Floor 和 Risk Ceiling 这两组枚举值最终要落回界面或日志，靠的是两个轻量取值函数。GetRiskFloor() 把 RiskFloorMin / RiskFloorMax / RiskFloorNone 分别映射成 "MINIMUM LOTSIZE"、"MAX-RISK"、"NONE"，未命中任何分支时返回空串。 天花板一侧更细一些：GetRiskCeil() 覆盖 RiskCeilMax 到 RiskCeilMax5 共 5 档，对应字符串 "MAX LOTSIZE" 至 "MAX LOTSIZE(x5)"，也就是允许把上限手数放大到经纪商最大手数的 1~5 倍。外汇与贵金属杠杆高，这种倍数放大在极端波动下可能迅速吞噬权益，实盘前务必在策略测试器里跑过。 构造函数把基础字段清成默认值：Lots=0.0、max_percent=100、ORDERTYPE 默认买单、OpenPrice 取当时 Ask()，ClosePrice 用 Ask()*1.01 再 NormalizePrice 归一。注意 ClosePrice 这个 1% 偏移只是初始化占位，并非真实平仓价假设。 Volume() 函数开头按 RiskProfileOption 做分支，具体手数算法在后续段落展开，当前片段只亮了 switch 入口。
class="kw">break; } } class=class="str">"cmt">//-- Retrieve user&class="macro">#x27;s Risk Floor Option class="type">class="kw">string GetRiskFloor() { class="kw">switch(RiskFloorOption) { case RiskFloorMin:class=class="str">"cmt">//MINIMUM LOTSIZE for Risk floor options class="kw">return "MINIMUM LOTSIZE"; class="kw">break; case RiskFloorMax:class=class="str">"cmt">//MAX-RISK for Risk floor options class="kw">return "MAX-RISK"; class="kw">break; case RiskFloorNone:class=class="str">"cmt">//NONE for Risk floor options class="kw">return "NONE"; class="kw">break; class="kw">default: class="kw">return ""; class="kw">break; } } class=class="str">"cmt">//-- Retrieve user&class="macro">#x27;s Risk Ceiling option class="type">class="kw">string GetRiskCeil() { class="kw">switch(RiskCeilOption) { case RiskCeilMax:class=class="str">"cmt">//MAX LOTSIZE for Risk ceiling options class="kw">return "MAX LOTSIZE"; class="kw">break; case RiskCeilMax2:class=class="str">"cmt">//MAX LOTSIZE(x2) for Risk ceiling options class="kw">return "MAX LOTSIZE(x2)"; class="kw">break; case RiskCeilMax3:class=class="str">"cmt">//MAX LOTSIZE(x3) for Risk ceiling options class="kw">return "MAX LOTSIZE(x3)"; class="kw">break; case RiskCeilMax4:class=class="str">"cmt">//MAX LOTSIZE(x4) for Risk ceiling options class="kw">return "MAX LOTSIZE(x4)"; class="kw">break; case RiskCeilMax5:class=class="str">"cmt">//MAX LOTSIZE(x5) for Risk ceiling options class="kw">return "MAX LOTSIZE(x5)"; class="kw">break; class="kw">default: class="kw">return ""; class="kw">break; } } class=class="str">"cmt">//Initialize values CRiskManagement::CRiskManagement(class="type">void):Lots(class="num">0.0),max_percent(class="num">100), ORDERTYPE(ORDER_TYPE_BUY),OpenPrice(Ask()), ClosePrice(NormalizePrice(Ask()+Ask()*class="num">0.01)) { } class="type">class="kw">double CRiskManagement::Volume() { class="kw">switch(RiskProfileOption) {
手数决策的 switch 分支与风险映射
这段 MT5 风控模块用 switch 把十种手数模式摊开：从经纪商最小手数 MINIMUM_LOT、最大手数 MAXIMUM_LOT，到按余额/空闲保证金百分比、固定金额、固定手数比例，再到自定义手数与最大风险百分比，各自返回或算出 Lots。 default 分支直接给 0.0 手数，意味着未识别的风险配置不会下单，相当于一道静默熔断。算完统一走 ValidateLotsize 与 NormalizeLotsize，把越界值拉回品种允许区间并规整小数位。 GetMinimumRisk 借 OrderCalcProfit 用最小手数反推最小风险金额；GetRisk 则按目标金额与最小金额的比值放大最小手数。外汇与贵金属杠杆波动大，这套映射若参数错配，可能瞬间吃掉空闲保证金，开 MT5 把 Risk_Profile 系列变量打印出来核对再实盘。
case MINIMUM_LOT:class=class="str">"cmt">//MINIMUM LOTSIZE - Risk Option class="kw">return LotsMin(); class="kw">break; case MAXIMUM_LOT:class=class="str">"cmt">//MAXIMUM LOTSIZE - Risk Option Lots = LotsMax(); class="kw">break; case PERCENTAGE_OF_BALANCE:class=class="str">"cmt">//PERCENTAGE OF BALANCE - Risk Option Lots = RiskProfile1(BALANCE); class="kw">break; case PERCENTAGE_OF_FREEMARGIN:class=class="str">"cmt">//PERCENTAGE OF FREE-MARGIN - Risk Option Lots = RiskProfile1(MARGIN); class="kw">break; case AMOUNT_PER_BALANCE:class=class="str">"cmt">//AMOUNT PER BALANCE - Risk Option Lots = RiskProfile2(BALANCE); class="kw">break; case AMOUNT_PER_FREEMARGIN:class=class="str">"cmt">//AMOUNT PER FREE-MARGIN - Risk Option Lots = RiskProfile2(MARGIN); class="kw">break; case LOTSIZE_PER_BALANCE:class=class="str">"cmt">//LOTSIZE PER BALANCE - Risk Option Lots = RiskProfile3(BALANCE); class="kw">break; case LOTSIZE_PER_FREEMARGIN:class=class="str">"cmt">//LOTSIZE PER FREE-MARGIN - Risk Option Lots = RiskProfile3(MARGIN); class="kw">break; case CUSTOM_LOT:class=class="str">"cmt">//CUSTOM LOTSIZE - Risk Option Lots = Risk_Profile_4; class="kw">break; case PERCENTAGE_OF_MAXRISK:class=class="str">"cmt">//PERCENTAGE OF MAX-RISK - Risk Option Lots = MaxRisk(Risk_Profile_5); class="kw">break; class="kw">default: Lots = class="num">0.0; class="kw">break; } ValidateLotsize(Lots);class=class="str">"cmt">//Check/Adjust Lotsize Limits NormalizeLotsize(Lots);class=class="str">"cmt">//Normalize Lotsize class="kw">return Lots; } class=class="str">"cmt">//-- Set Medium variable value class="type">void SetMedium(const RiskMedium R_Medium) {Medium = (R_Medium==BALANCE)?Account.Balance():Account.FreeMargin();} class=class="str">"cmt">//-- Get Minimum Risk for a Trade using Minimum Lot-size class="type">bool GetMinimumRisk() { class="kw">return OrderCalcProfit(ORDERTYPE,Symbol(),LotsMin(),OpenPrice,ClosePrice,MinimumAmount); } class=class="str">"cmt">//-- Retrieve Risk amount based on Risk inputs class="type">class="kw">double GetRisk(class="type">class="kw">double Amount) { if(!GetMinimumRisk()||Amount==class="num">0) class="kw">return class="num">0.0; class="kw">return ((Amount/MinimumAmount)*LotsMin()); } class=class="str">"cmt">//-- calculations for Risk options(PERCENTAGE OF BALANCE and PERCENTAGE OF FREE-MARGIN)
「三种风险档位与手数下限校验」
EA 里的仓位计算常按风险档位分流。下面这三段分别对应「按余额百分比」「按余额/空闲保证金比例」「按每单位余额手数」三种算法，跑通后直接回传风险金额或手数。 double CRiskManagement::RiskProfile1(const RiskMedium R_Medium) 先写入 Medium，再用 Risk_Profile_1 的百分比（除以 100）乘出 RiskAmount，最后 GetRisk 回传。若 Risk_Profile_1 设为 2，余额基数 10000，则风险额约 200。 RiskProfile2 里引入了 RiskAmountBoF 与 RiskAmount 的比值，小于 1 时强制取 1，相当于给空闲保证金模式加了下限保护；Medium<=0 直接返 0.0，避免除零或负仓。 RiskProfile3 更直白：Medium>0 时用手数系数 RiskLot 与 RiskLotBoF 推算，否则返 0.0。 ValidateLotsize 用手数引用做地板约束。RiskFloorMin 下，小于 LotsMin() 或大于 MaxRisk(max_percent) 一律压回最小手数；RiskFloorMax 允许顶到 MaxRisk(RiskFloorPercentage)，但仍托底 LotsMin()；RiskFloorNone 则低于最小手数直接归零——外汇与贵金属杠杆高，这种硬下限能避免奇点手数把账户拖进强平区。
class="type">class="kw">double CRiskManagement::RiskProfile1(const RiskMedium R_Medium) { SetMedium(R_Medium); RiskAmount = Medium*(Risk_Profile_1/class="num">100); class="kw">return GetRisk(RiskAmount); } class=class="str">"cmt">//-- calculations for Risk options(AMOUNT PER BALANCE and AMOUNT PER FREE-MARGIN) class="type">class="kw">double CRiskManagement::RiskProfile2(const RiskMedium R_Medium) { SetMedium(R_Medium); class="type">class="kw">double risk = (Risk_Profile_2.RiskAmountBoF/Risk_Profile_2.RiskAmount); risk = (risk<class="num">1)?class="num">1:risk; if(Medium<=class="num">0) class="kw">return class="num">0.0; RiskAmount = Medium/risk; class="kw">return GetRisk(RiskAmount); } class=class="str">"cmt">//-- calculations for Risk options(LOTSIZE PER BALANCE and LOTSIZE PER FREE-MARGIN) class="type">class="kw">double CRiskManagement::RiskProfile3(const RiskMedium R_Medium) { SetMedium(R_Medium); class="kw">return (Medium>class="num">0)?((Medium/Risk_Profile_3.RiskLotBoF)*Risk_Profile_3.RiskLot):class="num">0.0; } class="type">void CRiskManagement::ValidateLotsize(class="type">class="kw">double &Lotsize) { class="kw">switch(RiskFloorOption) { case RiskFloorMin:class=class="str">"cmt">//MINIMUM LOTSIZE for Risk floor options class=class="str">"cmt">//-- Check if lot-size is not less than Minimum lot or more than maximum allowable risk if(Lotsize<LotsMin()||Lotsize>MaxRisk(max_percent)) { Lotsize=LotsMin(); } class="kw">break; case RiskFloorMax:class=class="str">"cmt">//MAX-RISK for Risk floor options class=class="str">"cmt">//-- Check if lot-size is more the maximum allowable risk if(Lotsize>MaxRisk(max_percent)) { Lotsize=(MaxRisk(RiskFloorPercentage)>LotsMin())?MaxRisk(RiskFloorPercentage):LotsMin(); } else if(Lotsize<LotsMin())class=class="str">"cmt">//Check if lot-size is less than Minimum lot { Lotsize=LotsMin(); } class="kw">break; case RiskFloorNone:class=class="str">"cmt">//NONE for Risk floor options class=class="str">"cmt">//Check if lot-size is less than Minimum lot if(Lotsize<LotsMin()) { Lotsize=class="num">0.0; } class="kw">break; class="kw">default: Lotsize=class="num">0.0; class="kw">break; } class="kw">switch(RiskCeilOption) {
◍ 手数封顶与品种限仓的兜底逻辑
风险天花板那组 case 干的事很直白：算出理论手数后，按 RiskCeilMax 到 RiskCeilMax5 分别把上限锁在 LotsMax() 的 1~5 倍。比如 RiskCeilMax3 下，若 Lotsize 超过 LotsMax()*3 就直接砍回该值，避免单一信号在流动性差的时段把仓位顶到券商边界。 NormalizeLotsize 才是真正碰品种规则的环节。LotsLimit() 大于 0 时，它把已持仓 volume 加挂单 volume 再加本次 Lots，和限额比对；超了就反推 remaining_avail_lots = LotsLimit() - (PositionsVolume()+OrdersVolume())。 剩下可开手数若还大于等于 LotsMin()，且 RiskFloorOption 设成 RiskFloorMin，终端会打印“Warning: Volume Limit Reached, minimum Lotsize selected.”，实际下单倾向落到最小手数而非报错退出。外汇与贵金属杠杆波动大，这类限仓兜底能降低拒单概率，但无法消除滑点与极端行情风险，开 MT5 把 LotsLimit 打印出来核对自己的品种上限即可。
case RiskCeilMax:class=class="str">"cmt">//MAX LOTSIZE for Risk ceiling options class=class="str">"cmt">//Check if lot-size is more than Maximum lot if(Lotsize>LotsMax()) Lotsize=LotsMax(); class="kw">break; case RiskCeilMax2:class=class="str">"cmt">//MAX LOTSIZE(x2) for Risk ceiling options class=class="str">"cmt">//Check if lot-size is more than Maximum lot times two if(Lotsize>(LotsMax()*class="num">2)) Lotsize=(LotsMax()*class="num">2); class="kw">break; case RiskCeilMax3:class=class="str">"cmt">//MAX LOTSIZE(x3) for Risk ceiling options class=class="str">"cmt">//Check if lot-size is more than Maximum lot times three if(Lotsize>(LotsMax()*class="num">3)) Lotsize=(LotsMax()*class="num">3); class="kw">break; case RiskCeilMax4:class=class="str">"cmt">//MAX LOTSIZE(x4) for Risk ceiling options class=class="str">"cmt">//Check if lot-size is more than Maximum lot times four if(Lotsize>(LotsMax()*class="num">4)) Lotsize=(LotsMax()*class="num">4); class="kw">break; case RiskCeilMax5:class=class="str">"cmt">//MAX LOTSIZE(x5) for Risk ceiling options class=class="str">"cmt">//Check if lot-size is more than Maximum lot times five if(Lotsize>(LotsMax()*class="num">5)) Lotsize=(LotsMax()*class="num">5); class="kw">break; class="kw">default: class="kw">break; } } class="type">void CRiskManagement::NormalizeLotsize(class="type">class="kw">double &Lotsize) { if(Lotsize<=class="num">0.0) class="kw">return; class=class="str">"cmt">//-- Check if the is a Volume limit for the current Symbol if(LotsLimit()>class="num">0.0) { if((Lots+PositionsVolume()+OrdersVolume())>LotsLimit()) { class=class="str">"cmt">//-- calculation of available lotsize remaining class="type">class="kw">double remaining_avail_lots = (LotsLimit()-(PositionsVolume()+OrdersVolume())); if(remaining_avail_lots>=LotsMin()) { if(RiskFloorOption==RiskFloorMin)class=class="str">"cmt">//Check if Risk floor option is MINIMUM LOTSIZE { Print("Warning: Volume Limit Reached, minimum Lotsize selected.");
手数触顶时的降级与清零逻辑
这段手数计算收尾代码处理的是账户或品种成交量上限被碰触后的兜底动作。外汇与贵金属属高杠杆品种，触限本身就意味着风险敞口已逼近通道边界，必须让 EA 在超限时自动收敛而非硬挤单。 当风险下限模式设为 MAX-RISK 且剩余可用手数不足时，程序先打印“Volume Limit Reached, Lotsize Reduced.”警告，再把手数压到 remaining_avail_lots 与 RiskFloorPercentage 的乘积和 LotsMin() 之间取大值，保证不低于最小手数。 若连最小手数都分配不出来，直接打印“Volume Limit Reached!”并将 Lotsize 置 0、return 退出，本周期不再发单。最后一段用 LotsStep()>0.0 做守卫，把 LotSize 按步长向下取整：Lotsize = LotsStep() * MathFloor(Lotsize / LotsStep())，避免券商拒单。 开 MT5 把这段贴进自己的手数函数，改 RiskFloorPercentage 从 10 到 50 各跑一次历史回放，能直观看到触限日报里手数被砍的频次。
Lotsize = LotsMin(); } else if(RiskFloorOption==RiskFloorMax)class=class="str">"cmt">//Check if Risk floor option is MAX-RISK { Print("Warning: Volume Limit Reached, Lotsize Reduced."); Lotsize = ((remaining_avail_lots*(RiskFloorPercentage/class="num">100))>LotsMin())? (remaining_avail_lots*(RiskFloorPercentage/class="num">100)):LotsMin(); } } else { Print("Volume Limit Reached!"); Lotsize=class="num">0.0; class="kw">return; } } } class=class="str">"cmt">//Check if there is a valid Volume Step for the current Symbol if(LotsStep()>class="num">0.0) Lotsize=LotsStep()*MathFloor(Lotsize/LotsStep()); }
「通用图形类如何接管品种属性与风险面板」
在自建 MT5 面板时，CommonGraphics 类充当了“总装层”：它从 CObjectProperties、CChartProperties、CSymbolProperties 一路继承，把图表对象、品种属性、甚至部分风险管理接口揉进同一个实例里。这样做的好处是，你只要在构造函数里调一次 GraphicsRefresh，界面上的基础框体就会随品种切换自动重绘。 从继承链看，CSymbolProperties 与 CSymbolInfo 提供的点差是实时可读的，ContractSize() 返回的合约乘数直接决定你每手占用的保证金规模——外汇与贵金属杠杆高，错看合约尺寸可能让仓位失控。 GraphicsRefresh 里先画一个左上角坐标 (2,20) 到 (330,183) 的矩形作为底框，再叠 Symbol Name、Contract Size 等文本对象。注意 ContractSize() 取自品种属性，若你的经纪商对 XAUUSD 设的是 100 盎司/手，面板就会显示 100，而不是标准外汇的 100000。 别把继承当黑盒 继承层级深了以后，CRiskManagement 对象藏在 private 里，你改风险参数得确认它是否被 GraphicsRefresh 间接调用，否则面板显示和实际风控会脱节。
class="macro">#include "ObjectProperties.mqh" class="macro">#include "RiskManagement.mqh" class CCommonGraphics:CObjectProperties { class="kw">private: CRiskManagement CRisk;class=class="str">"cmt">//Risk management class object class="kw">public: CCommonGraphics(class="type">void);class=class="str">"cmt">//class constructor ~CCommonGraphics(class="type">void) {}\n class="type">void GraphicsRefresh();class=class="str">"cmt">//will create the chart objects }; CCommonGraphics::CCommonGraphics(class="type">void) { GraphicsRefresh();class=class="str">"cmt">//calling GraphicsRefresh function } class="type">void CCommonGraphics::GraphicsRefresh() { class=class="str">"cmt">//-- Will create the rectangle object Square(class="num">0,"Symbol Properties",class="num">2,class="num">20,class="num">330,class="num">183,ANCHOR_LEFT_UPPER); class=class="str">"cmt">//-- Will create the text object for the Symbol&class="macro">#x27;s name TextObj(class="num">0,"Symbol Name",Symbol(),class="num">5,class="num">23); class=class="str">"cmt">//-- Will create the text object for the contract size TextObj(class="num">0,"Symbol Contract Size","Contract Size: "+class="type">class="kw">string(ContractSize()),class="num">5,class="num">40,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the Symbol&class="macro">#x27;s Minimum lotsize
◍ 把品种手数约束画到图表左上角
想在 MT5 图表上实时盯住当前品种的合约参数，不必每次去规格窗口翻。上面这段用 CCommonGraphics 封装，把最小手数、最大手数、步长、限额以及风控模块的上下限一次性刷到左上角区域。 TextObj 的第一个参数是图表 ID（0 即当前图），第二个是对象名（不能重名，否则刷新时被覆盖），第三个是显示字符串，后面跟坐标与字号。注意 LotsMin() 返回 double，必须用 string() 转成文本，否则编译报错。 GraphicsRefresh 里先画了一个 2×20 到 330×183 的矩形做底框，Y 坐标从 23 到 180 按 20 像素步进排布。外汇和贵金属杠杆高、手数步长各品种不同，EURUSD 可能步长 0.01，XAUUSD 常见 0.01 但 LotsLimit 可能为 0（表示无硬上限），实盘前务必在策略测试器里跑一遍确认显示值。 下面逐行拆关键调用： TextObj(0,"Symbol MinLot","Minimum Lot: "+string(LotsMin()),5,60,CORNER_LEFT_UPPER,9); // 0=当前图表；"Symbol MinLot"=对象名；拼接最小手数文本；X=5,Y=60；锚定左上；字号9 TextObj(0,"Symbol MaxLot","Max Lot: "+string(LotsMax()),5,80,CORNER_LEFT_UPPER,9); // 同理显示最大手数，Y下移20避免重叠 TextObj(0,"Symbol Volume Step","Volume Step: "+string(LotsStep()),5,100,CORNER_LEFT_UPPER,9); // 显示手数步长，下单量必须是步长整数倍 TextObj(0,"Symbol Volume Limit","Volume Limit: "+string(LotsLimit()),5,120,CORNER_LEFT_UPPER,9); // 显示限额，返回0代表经纪商未设硬上限 TextObj(0,"Risk Option","Risk Option: "+CRisk.GetRiskOption(),5,140,CORNER_LEFT_UPPER,9); // 从风控类取风险选项文本 TextObj(0,"Risk Floor","Risk Floor: "+CRisk.GetRiskFloor(),5,160,CORNER_LEFT_UPPER,9); // 风控下限 TextObj(0,"Risk Ceil","Risk Ceiling: "+CRisk.GetRiskCeil(),5,180,CORNER_LEFT_UPPER,9); // 风控上限，Y=180仍在矩形183高度内
TextObj(class="num">0,"Symbol MinLot","Minimum Lot: "+class="type">class="kw">string(LotsMin()),class="num">5,class="num">60,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the Symbol&class="macro">#x27;s Maximum lotsize TextObj(class="num">0,"Symbol MaxLot","Max Lot: "+class="type">class="kw">string(LotsMax()),class="num">5,class="num">80,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the Symbol&class="macro">#x27;s Volume Step TextObj(class="num">0,"Symbol Volume Step","Volume Step: "+class="type">class="kw">string(LotsStep()),class="num">5,class="num">100,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the Symbol&class="macro">#x27;s Volume Limit TextObj(class="num">0,"Symbol Volume Limit","Volume Limit: "+class="type">class="kw">string(LotsLimit()),class="num">5,class="num">120,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the trader&class="macro">#x27;s Risk Option TextObj(class="num">0,"Risk Option","Risk Option: "+CRisk.GetRiskOption(),class="num">5,class="num">140,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the trader&class="macro">#x27;s Risk Floor TextObj(class="num">0,"Risk Floor","Risk Floor: "+CRisk.GetRiskFloor(),class="num">5,class="num">160,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the trader&class="macro">#x27;s Risk Ceiling TextObj(class="num">0,"Risk Ceil","Risk Ceiling: "+CRisk.GetRiskCeil(),class="num">5,class="num">180,CORNER_LEFT_UPPER,class="num">9); } class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">void CCommonGraphics::GraphicsRefresh() { class=class="str">"cmt">//-- Will create the rectangle object Square(class="num">0,"Symbol Properties",class="num">2,class="num">20,class="num">330,class="num">183,ANCHOR_LEFT_UPPER); class=class="str">"cmt">//-- Will create the text object for the Symbol&class="macro">#x27;s name TextObj(class="num">0,"Symbol Name",Symbol(),class="num">5,class="num">23); class=class="str">"cmt">//-- Will create the text object for the contract size TextObj(class="num">0,"Symbol Contract Size","Contract Size: "+class="type">class="kw">string(ContractSize()),class="num">5,class="num">40,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the Symbol&class="macro">#x27;s Minimum lotsize TextObj(class="num">0,"Symbol MinLot","Minimum Lot: "+class="type">class="kw">string(LotsMin()),class="num">5,class="num">60,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the Symbol&class="macro">#x27;s Maximum lotsize TextObj(class="num">0,"Symbol MaxLot","Max Lot: "+class="type">class="kw">string(LotsMax()),class="num">5,class="num">80,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the Symbol&class="macro">#x27;s Volume Step TextObj(class="num">0,"Symbol Volume Step","Volume Step: "+class="type">class="kw">string(LotsStep()),class="num">5,class="num">100,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the Symbol&class="macro">#x27;s Volume Limit
把风控参数钉在图表左上角
想在 MT5 图表上实时看自己这单的风险边界，不必开面板翻设置，直接往左上角画文本对象最省事。下面这段把品种成交量上限、风险选项、风险下限与上限一口气排进坐标 (5,120) 到 (5,180) 的区域，间距 20 像素，字体大小 9，锚点统一用 CORNER_LEFT_UPPER。 调用 TextObj 时第一个参数填 0 表示挂到当前图表；第二参是对象名，重名会被覆盖，所以用了 Symbol Volume Limit、Risk Option、Risk Floor、Risk Ceil 四个独立 ID。第三参拿 string() 把数值或方法返回值拼成可读串，LotsLimit() 与 CRisk 的几个 getter 必须在本作用域可见，否则编译期就报错。 外汇与贵金属杠杆高，这类只读展示只帮你盯边界，不替代经纪商强平逻辑；实际仓位触发 Risk Floor 或 Risk Ceiling 时，盘面提示和账户风控可能不同步，以终端提醒为准。
TextObj(class="num">0,"Symbol Volume Limit","Volume Limit: "+class="type">class="kw">string(LotsLimit()),class="num">5,class="num">120,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the trader&class="macro">#x27;s Risk Option TextObj(class="num">0,"Risk Option","Risk Option: "+CRisk.GetRiskOption(),class="num">5,class="num">140,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the trader&class="macro">#x27;s Risk Floor TextObj(class="num">0,"Risk Floor","Risk Floor: "+CRisk.GetRiskFloor(),class="num">5,class="num">160,CORNER_LEFT_UPPER,class="num">9); class=class="str">"cmt">//-- Will create the text object for the trader&class="macro">#x27;s Risk Ceiling TextObj(class="num">0,"Risk Ceil","Risk Ceiling: "+CRisk.GetRiskCeil(),class="num">5,class="num">180,CORNER_LEFT_UPPER,class="num">9); }
「把新闻EA挂上图表前的初始化动作」
这套新闻交易EA在编译通过后不会立刻下单，所有逻辑要等它被拖到图表上才启动。你先选好想监控的交易品种窗口，在附加EA之前，图表形态大致和官方示例图一致，但背景色由经纪商主题决定，可能出现差异。 首次运行NewsTrading 2.00且通用文件夹里没有旧版1.00建立的日历库时，EA会自行初始化数据库，图表会按默认样式显示。若终端连不上数据源，日历拉取会失败，手数也不会按风险选项计算出来。 如果机器上跑过1.00版，2.00首次加载会直接清掉旧版的所有表格再重建，避免字段冲突。风险管理输入变量可以先留默认值，下面这段头文件定义了画布尺寸和风控参数分组，值得逐行看。 代码里的IMG_WIDTH和IMG_HEIGHT写死为200与100，表示用于绘制面板的离屏缓冲大小；RISK_Mini_Percent默认75代表在最小手数模式下最多动用账户75%的风险阈值，Risk_2_VALUE=1000配合Risk_2_AMOUNT=10意为每1000单位净值拨10元风险额。外汇与贵金属杠杆高，实盘加载前建议在模拟盘验证这套映射。
class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| NewsTrading | class=class="str">"cmt">//| Copyright class="num">2024, MetaQuotes Ltd. | class=class="str">"cmt">//| [MQL5官方文档] | class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//--- width and height of the canvas(used for drawing) class="macro">#define IMG_WIDTH class="num">200 class="macro">#define IMG_HEIGHT class="num">100 class=class="str">"cmt">//--- enable to set class="type">class="kw">color format ENUM_COLOR_FORMAT clr_format=COLOR_FORMAT_XRGB_NOALPHA; class=class="str">"cmt">//--- drawing array(buffer) class="type">uint ExtImg[IMG_WIDTH*IMG_HEIGHT]; class="macro">#include "News.mqh" CNews NewsObject;class=class="str">"cmt">//Class CNews Object &class="macro">#x27;NewsObject&class="macro">#x27; class="macro">#include "TimeManagement.mqh" CTimeManagement CTM;class=class="str">"cmt">//Class CTimeManagement Object &class="macro">#x27;CTM&class="macro">#x27; class="macro">#include "WorkingWithFolders.mqh" CFolders Folder();class=class="str">"cmt">//Calling Class&class="macro">#x27;s Constructor class="macro">#include "ChartProperties.mqh" CChartProperties CChart;class=class="str">"cmt">//Class CChartProperties Object &class="macro">#x27;CChart&class="macro">#x27; class="macro">#include "RiskManagement.mqh" CRiskManagement CRisk;class=class="str">"cmt">//Class CRiskManagement Object &class="macro">#x27;CRisk&class="macro">#x27; class="macro">#include "CommonGraphics.mqh" CCommonGraphics CGraphics();class=class="str">"cmt">//Calling Class&class="macro">#x27;s Constructor enum iSeparator { Delimiterclass=class="str">"cmt">//__________________________ }; sinput group "+--------| RISK MANAGEMENT |--------+"; input RiskOptions RISK_Type=MINIMUM_LOT;class=class="str">"cmt">//SELECT RISK OPTION input RiskFloor RISK_Mini=RiskFloorMin;class=class="str">"cmt">//RISK FLOOR input class="type">class="kw">double RISK_Mini_Percent=class="num">75;class=class="str">"cmt">//MAX-RISK [class="num">100<-->class="num">0.01]% input RiskCeil RISK_Maxi=RiskCeilMax;class=class="str">"cmt">//RISK CEILING sinput iSeparator iRisk_1=Delimiter;class=class="str">"cmt">//__________________________ sinput iSeparator iRisk_1L=Delimiter;class=class="str">"cmt">//PERCENTAGE OF [BALANCE | FREE-MARGIN] input class="type">class="kw">double Risk_1_PERCENTAGE=class="num">3;class=class="str">"cmt">//[class="num">100<-->class="num">0.01]% sinput iSeparator iRisk_2=Delimiter;class=class="str">"cmt">//__________________________ sinput iSeparator iRisk_2L=Delimiter;class=class="str">"cmt">//AMOUNT PER [BALANCE | FREE-MARGIN] input class="type">class="kw">double Risk_2_VALUE=class="num">1000;class=class="str">"cmt">//[BALANCE | FREE-MARGIN] input class="type">class="kw">double Risk_2_AMOUNT=class="num">10;class=class="str">"cmt">//EACH AMOUNT sinput iSeparator iRisk_3=Delimiter;class=class="str">"cmt">//__________________________ sinput iSeparator iRisk_3L=Delimiter;class=class="str">"cmt">//LOTSIZE PER [BALANCE | FREE-MARGIN] input class="type">class="kw">double Risk_3_VALUE=class="num">1000;class=class="str">"cmt">//[BALANCE | FREE-MARGIN] input class="type">class="kw">double Risk_3_LOTSIZE=class="num">0.1;class=class="str">"cmt">//EACH LOTS(VOLUME) sinput iSeparator iRisk_4=Delimiter;class=class="str">"cmt">//__________________________ sinput iSeparator iRisk_4L=Delimiter;class=class="str">"cmt">//CUSTOM LOTSIZE input class="type">class="kw">double Risk_4_LOTSIZE=class="num">0.01;class=class="str">"cmt">//LOTS(VOLUME) sinput iSeparator iRisk_5=Delimiter;class=class="str">"cmt">//__________________________
◍ 风控档位的初始化边界
EA 启动时把外部输入的风险参数灌进 CRiskManagement 类，这里最容易被忽略的是边界钳制：任何百分比档位若超出 100% 或低于 0.01%，都会被强制拉回区间内。比如 Risk_5_PERCENTAGE 默认填 1，代表占用最大风险的 1%，但用户手滑填 200 也只会按 100 处理。 初始化段里 RiskFloorPercentage 同样走这套逻辑，RISK_Mini_Percent 低于 0.01 按 0.01 算、高于 100 按 100 算，避免地板风险变成负数或满仓。五个风险剖面（百分比、金额、手数、固定手、最大风险占比）各自独立赋值，互不覆盖。 非回测环境下，OnInit 还会建一个 OBJ_BITMAP_LABEL 名为 STATUS，锚点设在距左 5 像素、距顶 22 像素，绑定图形资源 ::PROGRESS 用来画状态位图。开 MT5 把这段塞进自己 EA 的 OnInit，改掉输入参数名就能直接复用边界钳制写法。
sinput iSeparator iRisk_5L=Delimiter;class=class="str">"cmt">//PERCENTAGE OF MAX-RISK input class="type">class="kw">double Risk_5_PERCENTAGE=class="num">1;class=class="str">"cmt">//[class="num">100<-->class="num">0.01]% class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| Expert initialization function | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">int OnInit() { class=class="str">"cmt">//--- class=class="str">"cmt">//Initializing CRiskManagement variable for Risk options RiskProfileOption = RISK_Type; class=class="str">"cmt">//Initializing CRiskManagement variable for Risk floor RiskFloorOption = RISK_Mini; class=class="str">"cmt">//Initializing CRiskManagement variable for RiskFloorMax RiskFloorPercentage = (RISK_Mini_Percent>class="num">100)?class="num">100: (RISK_Mini_Percent<class="num">0.01)?class="num">0.01:RISK_Mini_Percent;class=class="str">"cmt">//Percentage cannot be more than class="num">100% or less than class="num">0.01% class=class="str">"cmt">//Initializing CRiskManagement variable for Risk ceiling RiskCeilOption = RISK_Maxi; class=class="str">"cmt">//Initializing CRiskManagement variable for Risk options(PERCENTAGE OF BALANCE and PERCENTAGE OF FREE-MARGIN) Risk_Profile_1 = (Risk_1_PERCENTAGE>class="num">100)?class="num">100: (Risk_1_PERCENTAGE<class="num">0.01)?class="num">0.01:Risk_1_PERCENTAGE;class=class="str">"cmt">//Percentage cannot be more than class="num">100% or less than class="num">0.01% class=class="str">"cmt">//Initializing CRiskManagement variables for Risk options(AMOUNT PER BALANCE and AMOUNT PER FREE-MARGIN) Risk_Profile_2.RiskAmountBoF = Risk_2_VALUE; Risk_Profile_2.RiskAmount = Risk_2_AMOUNT; class=class="str">"cmt">//Initializing CRiskManagement variables for Risk options(LOTSIZE PER BALANCE and LOTSIZE PER FREE-MARGIN) Risk_Profile_3.RiskLotBoF = Risk_3_VALUE; Risk_Profile_3.RiskLot = Risk_3_LOTSIZE; class=class="str">"cmt">//Initializing CRiskManagement variable for Risk option(CUSTOM LOTSIZE) Risk_Profile_4 = Risk_4_LOTSIZE; class=class="str">"cmt">//Initializing CRiskManagement variable for Risk option(PERCENTAGE OF MAX-RISK) Risk_Profile_5 = (Risk_5_PERCENTAGE>class="num">100)?class="num">100: (Risk_5_PERCENTAGE<class="num">0.01)?class="num">0.01:Risk_5_PERCENTAGE;class=class="str">"cmt">//Percentage cannot be more than class="num">100% or less than class="num">0.01% CChart.ChartRefresh();class=class="str">"cmt">//Load chart configurations CGraphics.GraphicsRefresh();class=class="str">"cmt">//-- Create/Re-create chart objects if(!MQLInfoInteger(MQL_TESTER))class=class="str">"cmt">//Checks whether the program is in the strategy tester { class=class="str">"cmt">//--- create OBJ_BITMAP_LABEL object for drawing ObjectCreate(class="num">0,"STATUS",OBJ_BITMAP_LABEL,class="num">0,class="num">0,class="num">0); ObjectSetInteger(class="num">0,"STATUS",OBJPROP_XDISTANCE,class="num">5); ObjectSetInteger(class="num">0,"STATUS",OBJPROP_YDISTANCE,class="num">22); class=class="str">"cmt">//--- specify the name of the graphical resource ObjectSetString(class="num">0,"STATUS",OBJPROP_BMPFILE,"::PROGRESS"); class="type">uint w,h; class=class="str">"cmt">// variables for receiving text class="type">class="kw">string sizes class="type">uint x,y; class=class="str">"cmt">// variables for calculation of the current coordinates of text class="type">class="kw">string anchor points /* In the Do while loop below, the code will check if the terminal is connected to the internet. If the the program is stopped the loop will class="kw">break, if the program is not stopped and the terminal
断网时的等待动画怎么画
在 MT5 指标或 EA 的 OnDraw 类逻辑里，若终端未连网，程序不能干等，得给盘面留个可见信号。上面这段代码用 do-while 循环配合静态变量 dots，在离线态下不断往 ExtImg 缓冲画 'Waiting' 加渐增的小圆点，形成加载动画。 关键点在于：每次循环先用 ArrayFill 把 IMG_WIDTH*IMG_HEIGHT 长度的 ExtImg 清零，避免上一帧残影；TextSetFont 设 Arial 负字号（-150 表示 15pt 且按像素映射）与 FW_EXTRABOLD 粗体，再用 TextOut 把文字以 TA_LEFT|TA_TOP 对齐写进缓冲。 坐标计算上，'Waiting' 水平固定 x=10，垂直取 IMG_HEIGHT/2 - h/2 做居中；点号区从 w+13 起笔（w 是 'Waiting' 实测像素宽），用 StringSubstr("...",0,dots) 截 0~3 个点循环。最后 ResourceCreate("::PROGRESS",...) 更新图形资源并 ChartRedraw() 强制重绘，Sleep(500) 控节奏——这意味着动画每秒约刷新 2 次。 这种写法直接可抄进你自己的离线提示模块；外汇与贵金属行情依赖实时连接，断线后任何自动策略都该有显式等待态，否则容易误判。
class="type">bool done=false; do { class=class="str">"cmt">//--- clear the drawing buffer array ArrayFill(ExtImg,class="num">0,IMG_WIDTH*IMG_HEIGHT,class="num">0); if(!TerminalInfoInteger(TERMINAL_CONNECTED)) { class=class="str">"cmt">//-- integer dots used as a loading animation class="kw">static class="type">int dots=class="num">0; class=class="str">"cmt">//--- set the font TextSetFont("Arial",-class="num">150,FW_EXTRABOLD,class="num">0); TextGetSize("Waiting",w,h);class=class="str">"cmt">//get text width and height values class=class="str">"cmt">//--- calculate the coordinates of the &class="macro">#x27;Waiting&class="macro">#x27; text x=class="num">10;class=class="str">"cmt">//horizontal alignment y=IMG_HEIGHT/class="num">2-(h/class="num">2);class=class="str">"cmt">//alignment for the text to be centered vertically class=class="str">"cmt">//--- output the &class="macro">#x27;Waiting&class="macro">#x27; text to ExtImg[] buffer TextOut("Waiting",x,y,TA_LEFT|TA_TOP,ExtImg,IMG_WIDTH,IMG_HEIGHT,ColorToARGB(CChart.SymbolBackground()),clr_format); class=class="str">"cmt">//--- calculate the coordinates for the dots after the &class="macro">#x27;Waiting&class="macro">#x27; text x=w+class="num">13;class=class="str">"cmt">//horizontal alignment y=IMG_HEIGHT/class="num">2-(h/class="num">2);class=class="str">"cmt">//alignment for the text to be centered vertically TextSetFont("Arial",-class="num">160,FW_EXTRABOLD,class="num">0); class=class="str">"cmt">//--- output of dots to ExtImg[] buffer TextOut(StringSubstr("...",class="num">0,dots),x,y,TA_LEFT|TA_TOP,ExtImg,IMG_WIDTH,IMG_HEIGHT,ColorToARGB(CChart.SymbolBackground()),clr_format); class=class="str">"cmt">//--- update the graphical resource ResourceCreate("::PROGRESS",ExtImg,IMG_WIDTH,IMG_HEIGHT,class="num">0,class="num">0,IMG_WIDTH,clr_format); class=class="str">"cmt">//--- force chart update ChartRedraw(); dots=(dots==class="num">3)?class="num">0:dots+class="num">1; class=class="str">"cmt">//-- Notify user that program is waiting for connection Print("Waiting for connection..."); Sleep(class="num">500); class="kw">continue; } else {
「初始化时的数据库就绪检查」
在指标或EA初始化阶段，若检测到新闻数据库尚未建立，程序会先画一张居中显示“Getting Ready...”的进度图，再循环等待连接成功。连接一旦成立，终端会打印“Connection Successful!”并调用 CreateEconomicDatabase() 建库，随后删掉名为“STATUS”的图表对象并强制重绘。 下面这段片段演示了字体设定与文字输出的关键调用：TextSetFont 用 Arial、字号 -120、FW_EXTRABOLD 粗体；TextGetSize 取“Getting Ready”的宽高，便于按 IMG_HEIGHT/2 - h/2 做垂直居中。ResourceCreate 把 ExtImg[] 缓冲注册成“::PROGRESS”资源，ChartRedraw 立刻刷新。 若程序跑在策略测试器里而必要文件不存在，初始化会直接返回 INIT_FAILED，并打印“Run Program outside of the Strategy Tester”。实盘或普通图表环境下，FileIsExist 查不到 NEWS_DATABASE_FILE 也会同样失败。 另一个坑在回测数据时效性：用 NewsObject.GetLatestNewsDate() 加一天得出 latestdate，若它小于 TimeCurrent()，说明库已过期，之后日期在回测中将无数据可用。终端会打出“Database Dates End at:”及具体时间，提醒你先脱开测试器跑一次更新。外汇与贵金属消息面驱动大，数据库缺漏可能让回测结果偏离实盘，属高风险项。
class=class="str">"cmt">//--- set the font TextSetFont("Arial",-class="num">120,FW_EXTRABOLD,class="num">0); TextGetSize("Getting Ready",w,h);class=class="str">"cmt">//get text width and height values x=class="num">20;class=class="str">"cmt">//horizontal alignment y=IMG_HEIGHT/class="num">2-(h/class="num">2);class=class="str">"cmt">//alignment for the text to be centered vertically class=class="str">"cmt">//--- output the text &class="macro">#x27;Getting Ready...&class="macro">#x27; to ExtImg[] buffer TextOut("Getting Ready...",x,y,TA_LEFT|TA_TOP,ExtImg,IMG_WIDTH,IMG_HEIGHT,ColorToARGB(CChart.SymbolBackground()),clr_format); class=class="str">"cmt">//--- update the graphical resource ResourceCreate("::PROGRESS",ExtImg,IMG_WIDTH,IMG_HEIGHT,class="num">0,class="num">0,IMG_WIDTH,clr_format); class=class="str">"cmt">//--- force chart update ChartRedraw(); class=class="str">"cmt">//-- Notify user that connection is successful Print("Connection Successful!"); NewsObject.CreateEconomicDatabase();class=class="str">"cmt">//calling the database create function done=true; } } while(!done&&!IsStopped()); class=class="str">"cmt">//-- Delete chart object ObjectDelete(class="num">0,"STATUS"); class=class="str">"cmt">//-- force chart to update ChartRedraw(); } else { class=class="str">"cmt">//Checks whether the database file exists if(!FileIsExist(NEWS_DATABASE_FILE,FILE_COMMON)) { Print("Necessary Files Do not Exist!"); Print("Run Program outside of the Strategy Tester"); Print("Necessary Files Should be Created First"); class="kw">return(INIT_FAILED); } else { class=class="str">"cmt">//Checks whether the lastest database date includes the time and date being tested class="type">class="kw">datetime latestdate = CTM.TimePlusOffset(NewsObject.GetLatestNewsDate(),CTM.DaysS());class=class="str">"cmt">//Day after the lastest recorded time in the database if(latestdate<TimeCurrent()) { Print("Necessary Files outdated!"); Print("To Update Files: Run Program outside of the Strategy Tester"); } Print("Database Dates End at: ",latestdate); PrintFormat("Dates after %s will not be available for backtest",TimeToString(latestdate));
◍ 手算仓位与风险类型在初始化里的落地
这段尾部代码处在 EA 的 OnInit 收口位置，前面已经跑完 CRisk 对象的仓位计算逻辑，这里只负责把结果吐到日志里供人工核对。
| Print("Lots: ",CRisk.Volume()," | Risk type: ",CRisk.GetRiskOption()); 这行把实际开仓手数和交易者选定的风险类型（固定、百分比或equity比例等）打印出来，回测或实盘前先在 MT5 终端的 Experts 标签里看一眼，能避免把测试参数当成实盘参数挂上去。 |
|---|
OnDeinit 和 OnTick 在这里都是空壳，说明该示例把全部风控初始化压在 OnInit 一次性完成；若你要在 tick 级动态改手数，得自己补 OnTick 里的重算调用，否则日志里的 Lots 不会随净值波动自动变。外汇与贵金属杠杆高，任何手数计算错误都可能放大穿仓概率，上线前务必用策略测试器跑一遍不同余额档位。
}
}
class=class="str">"cmt">//-- the volume calculations and the risk type set by the trader
Print("Lots: ",CRisk.Volume()," || Risk type: ",CRisk.GetRiskOption());
class=class="str">"cmt">//---
class="kw">return(INIT_SUCCEEDED);
}
class=class="str">"cmt">//+------------------------------------------------------------------+
class=class="str">"cmt">//| Expert deinitialization function |
class=class="str">"cmt">//+------------------------------------------------------------------+
class="type">void OnDeinit(const class="type">int reason)
{
class=class="str">"cmt">//---
}
class=class="str">"cmt">//+------------------------------------------------------------------+
class=class="str">"cmt">//| Expert tick function |
class=class="str">"cmt">//+------------------------------------------------------------------+
class="type">void OnTick()
{
class=class="str">"cmt">//---
}
class=class="str">"cmt">//+------------------------------------------------------------------+别急着下结论
这套新闻交易框架到目前只搭完了骨架：夏令时父类统管各时区子类，交易品种与图表属性各自独立成类，SQLite 读写也做了效率对比和简化。对象属性类负责图表对象的增删，风险管理类按不同偏好给仓位约束，通用图形类把品种属性和风险选项直接画到图上。 从社区反馈看，把止损切到 0 会触发异常，作者已在后续篇处理；另有交易者指出佣金和掉期必须计入盈亏平衡，否则自动系统显示的“盈利”可能只是毛利润。外汇与贵金属自带高杠杆风险，回测通过不等于实盘能跑通。 下一篇才真正碰开仓执行，当前这些类你都能在 NewsTrading_Part2_ProjectFiles.zip（440.93 KB）里直接编译验证。先把父类和子类的继承关系在 MT5 里跑顺，比急着等第三部分更有用。