Article ID: 244230 - Last Review: March 14, 2006 - Revision: 2.1

FIX: Slow Query Performance Using Correlated Subquery with "Or Exists" Clause

This article was previously published under Q244230
BUG #: 55001 (SQLBUG_70)
Expand all | Collapse all

SYMPTOMS

A query that contains a correlated subquery with an OR EXISTS clause may perform slowly.

WORKAROUND

You can rewrite the query and include a redundant clause in the subquery as shown in the following code example:
select A.col1 from TableA A, TableB B 
where 
A.col1 = B.col1
and B.col2 = 'X'
and  
(A.col2 = 'Y' 
  or  exists (select * from tableC C
                 where A.col3 = C.col3
		 -- Add redundant clause from outer query
                 and B.col2 = 'X'
                 )
)
				

STATUS

Microsoft has confirmed this to be a problem in SQL Server 7.0. This problem has been corrected in U.S. Service Pack 1 for Microsoft SQL Server 7.0. For more information, click the following article number to view the article in the Microsoft Knowledge Base:
232570  (http://support.microsoft.com/kb/232570/ ) INF: How to Obtain Service Pack 1 for Microsoft SQL Server 7.0 and Microsoft Data Engine (MSDE) 1.0
For more information, contact your primary support provider.

MORE INFORMATION

The following sample code demonstrates the problem:
select A.col1 from TableA A, TableB B 
where 
A.col1 = B.col1
and B.col2 = 'X'
and  
(A.col2 = 'Y' 
  or  exists (select * from tableC C
                 where A.col3 = C.col3
		 --and B.col2 = 'X'
                 )
)
				
The SHOWPLAN output for this query shows the optimizer is picking a hash match join strategy.

APPLIES TO
  • Microsoft SQL Server 7.0 Standard Edition
Keywords: 
kbbug kbfix kbqfe KB244230
 

Article Translations

 

Related Support Centers