Ø 简介
在平常编写增删改的 SQL 语句时，我们用的最多的就是 INSERT、UPDATE 和 DELETE 语句，这是最基本的增删改语句。其实，SQL Server 中还有另外一个可以实现增删改的语句，那就是本文将要介绍的 MERGE 语句。
1. 什么是 MERGE 语句
1) MERGE 语句可以在一条语句中，根据与源表联接的结果，对目标表执行插入、更新或删除操作。 例如，根据在另一个表中找到的差异在一个表中插入、更新或删除行，可以对两个表进行同步。
2) 为什么使用 MERGE 语句呢？
1. 试想一下，如果我们想要完成这件事情：将查询的结果集写入 A 表中，如果 A 表已经存在了该记录，就执行 UPDATE 语句（更新）；否则，执行 INSERT 语句（新增）。（删除也是如此）
2. 如果按照我们的思路，是不是首先应该查出数据，在根据关键字段去 A 表中匹配，如果匹配到则执行 UPDATE 语句，否则执行 INSERT 语句。但是这样一来我们必须执行两次查询结果集：第一次查询比对，第二次查询后执行新增或更新操作，这样就相当繁琐也影响性能。
3. 而 MERGE 语句可以很轻松的帮我们解决这个问题，帮我们简洁了代码和优化性能。
2. 我的示例
1) 使用 MERGE语句完成新增或更新操作
INSERT INTO Sys_DbLog(LogTime, Type, Itme, Message, Remark) VALUES(GETDATE(), 'info', 'SP_Update_AchievCommission', '存储过程执行', NULL);
BEGIN TRY
WITH ce1(OrderNum, TotalMoney) AS
(
SELECT OrderNum, SUM(T2.TotalMoney) AS TotalMoney FROM Orders AS T1
INNER JOIN OrderDetail AS T2 ON(T1.Id = T2.OrderId)
WHERE 1=1
AND (T1.OrderStatusId > 1 AND T1.OrderStatusId < 10)
AND (T1.PayStatusId = 2 AND T1.PayTime >= @StartTime AND T1.PayTime <= @EndTime)
GROUP BY T1.OrderNum
) --订单明细的总实付金额
MERGE INTO Crm_AchievCommission AS T1
USING
(
SELECT T1.OrderNum, T1.PayTime, T1.CityId AS OrderCityId, T3.CityId AS CustCityId, T4.CityId AS EmplCityId, T4.EmployeeId, T5.ProductId,
CAST(ROUND((T1.RealTotal - ISNULL(T1.Freight, 0)) * (ISNULL(T5.TotalMoney, 0) / T7.TotalMoney), 2) AS numeric(19,2)) AS RealTotal/*四舍五入保留两位小数*/,
T6.CommissionRate, T6.ExtraCommissionRate FROM Orders AS T1
INNER JOIN UserInfo AS T2 ON(T1.UserId = T2.Id)
INNER JOIN Customer AS T3 ON(T2.CustomerId = T3.Id)
INNER JOIN Sys_EmployeeProfile AS T4 ON(T1.SalesUserId = T4.EmployeeId)
INNER JOIN OrderDetail AS T5 ON(T1.Id = T5.OrderId)
INNER JOIN Crm_CommissionRates AS T6 ON(T5.ProductId = T6.ProductId)
INNER JOIN ce1 AS T7 ON(T1.OrderNum = T7.OrderNum)
WHERE 1=1
AND T1.UserId > 0
AND T2.CustomerId > 0
AND T1.SalesUserId > 0
AND (T1.OrderStatusId > 1 AND T1.OrderStatusId < 10)
AND (T1.PayStatusId = 2 AND T1.PayTime >= @StartTime AND T1.PayTime <= @EndTime)
) AS T2 ON(T1.OrderNum = T2.OrderNum AND T1.ProductId = T2.ProductId)
WHEN MATCHED
THEN UPDATE SET
OrderCityId = T2.OrderCityId,
CustCityId = T2.CustCityId,
EmplCityId = T2.EmplCityId,
EmployeeId = T2.EmployeeId,
RealTotal = T2.RealTotal,
UpdateTime = GETDATE()
WHEN NOT MATCHED
THEN INSERT(OrderNum, PayTime, OrderCityId, CustCityId, EmplCityId, EmployeeId, ProductId, RealTotal, CommissionRate, ExtraCommissionRate, CreateTime, UpdateTime)
VALUES(T2.OrderNum, T2.PayTime, T2.OrderCityId, T2.CustCityId, T2.EmplCityId, T2.EmployeeId, T2.ProductId, T2.RealTotal, T2.CommissionRate, T2.ExtraCommissionRate, GETDATE(), NULL);
END TRY
BEGIN CATCH
DECLARE @Message nvarchar(4000) = '错误行号：' + CAST(ERROR_LINE() AS nvarchar(20)) + '，错误信息：' + ERROR_MESSAGE();
INSERT INTO Sys_DbLog(LogTime, Type, Itme, Message, Remark) VALUES(GETDATE(), 'error', 'SP_Update_AchievCommission', @Message, NULL);
END CATCH
3. 更多介绍
1) 官网介绍
https://docs.microsoft.com/zh-cn/sql/t-sql/statements/merge-transact-sql?view=sql-server-2017
2) 其他