Question

该表是一张大表，数百万条记录。

表1有三栏

ID，列X和列Y.目前我正在使用＆＃34; in＆＃34;声明很慢;特别是在大表中。我希望在不使用in语句的情况下提高此更新语句的性能。非常感谢任何帮助

    Update [Table1] set Column_X='Delete' where Column_Y in(    
    select distinct (Column_Y) from [Table1]  where Column_X='Delete'
    )

   result before running script
   Table 1

   ID     Column X    Column Y      
   1      Delete      CAT           
   2      x1          CAT           
   3      x1          CAT        
   4      x1          COW           
   5      x1          COW     
   6      x2          Moon           
   7      Delete      Chicken         


   Intended result after running script
   Table 1

   ID     Column X    Column Y      
   1      Delete      CAT           
   2      Delete      CAT           
   3      Delete      CAT        
   4      x1          COW           
   5      x1          COW     
   6      x2          Moon           
   7      Delete      Chicken

Answer 1

您可以在此使用exists子句。

Update [Table1] set Column_X='Delete' from [Table1] t1where exists(
select t2.Column_Y
from [Table1] t2
where t2.Column_X = 'Delete'
and t2.Column_Y = t1.Column_Y)

它应该比“in”子句快一点。

P.S。：从不使用“in”，因为它进行全表扫描。 “存在”条款应该适用于几乎所有情况。

Answer 2

您可以在FROM子句中多次使用同一个表，例如：

update Table1
   set Column_X = 'Delete'
  from Table1 a, Table1 b
 where a.Column_Y = b.Column_Y
   and b.Column_X = 'Delete'

更新会影响Table1的第一个实例，即＆＃34; a＆＃34;。

更新BIG表中的记录而不使用＆＃39; in＆＃39;声明

2 个答案: