今天我用到了这样一组SQL语句，它所执行的表中有130万数据，

SELECT  COUNT(a.A0100) AS zzPsnNum, SUM(isnull(a.A5830,0)) AS ZzNums, Gz_Tz_B001.B0105 as DwName, Gz_Tz_B001.B0110 as B0110
FROM   (

 SELECT A0100, B0110,CurrWageMonth,A5830 FROM  Gz_Tz_Zz 
 UNION
 SELECT A0100, B0110,CurrWageMonth, A6808 AS A5830 FROM  Gz_Tz_Dgry  
 ) AS a
  RIGHT OUTER JOIN Gz_Tz_B001 ON a.B0110 = Gz_Tz_B001.B0110 AND a.CurrWageMonth = Gz_Tz_B001.currWageMonth
 WHERE  (Gz_Tz_B001.CurrWageMonth = '200806' And Gz_Tz_B001.B0110 like '037%'  And Gz_Tz_B001.Codeid='UN' AND (Gz_Tz_B001.B0143 = '1') )

 GROUP BY Gz_Tz_B001.B0110, Gz_Tz_B001.B0105,Gz_Tz_B001.OrderIDjg
 ORDER BY Gz_Tz_B001.OrderIDJg

执行完上面这条语句用了24分钟。

我把SQL语句加了条件，修改成下面这样后，速度竟然快了 几百倍

SELECT  COUNT(a.A0100) AS zzPsnNum, SUM(isnull(a.A5830,0)) AS ZzNums, Gz_Tz_B001.B0105 as DwName, Gz_Tz_B001.B0110 as B0110
FROM   (

 SELECT A0100, B0110,CurrWageMonth,A5830 FROM  Gz_Tz_Zz  where CurrWageMonth = '200806'
 UNION
 SELECT A0100, B0110,CurrWageMonth, A6808 AS A5830 FROM  Gz_Tz_Dgry  where CurrWageMonth = '200806'
 
 ) AS a
  RIGHT OUTER JOIN Gz_Tz_B001 ON a.B0110 = Gz_Tz_B001.B0110 AND a.CurrWageMonth = Gz_Tz_B001.currWageMonth
 WHERE  (Gz_Tz_B001.CurrWageMonth = '200806' And Gz_Tz_B001.B0110 like '037%'  And Gz_Tz_B001.Codeid='UN' AND (Gz_Tz_B001.B0143 = '1') )

 GROUP BY Gz_Tz_B001.B0110, Gz_Tz_B001.B0105,Gz_Tz_B001.OrderIDjg
 ORDER BY Gz_Tz_B001.OrderIDJg

其实就是让连接的左部分数据减少了，但是两条语句执行的结果一样的。