Comments on: SQL code made simpler

By: The Most-Watched IT Questions: July 5, 2011 - ITKE Community Blog

The Most-Watched IT Questions: July 5, 2011 - ITKE Community Blog — Tue, 05 Jul 2011 14:59:40 +0000

[...] 3. Carlosdl and TomLiotta help out a fellow member: SQL code made simpler. [...]

By: The Most-Watched IT Questions: June 28, 2011 - ITKE Community Blog

The Most-Watched IT Questions: June 28, 2011 - ITKE Community Blog — Tue, 28 Jun 2011 07:00:53 +0000

[...] 5. Carlosdl and TomLiotta try to help making some SQL code simpler. [...]

By: tomliotta

tomliotta — Fri, 24 Jun 2011 01:33:56 +0000

I want to update the same column (tSTAdm) in File_A when it is not equal to the same column (tStAdm) in File_BU.

Because of “a unique key – TsStat“, I assume you mean that you simply want File_a and File_BU to have matching tStADM values, based on matches for TsStat.

But you don’t want to UPDATE all rows because many rows already have correct (i.e., matching) values.

I’d say Carlosdl has the right answer. And if performance is the main reason to avoid just doing a straight UPDATE, then it’s probably the best answer.

Tom

By: carlosdl

carlosdl — Thu, 23 Jun 2011 20:43:36 +0000

This is not simpler, but it could be faster in some circumstances:

UPDATE File_a as T01 SET T01.tStADM = 
	(Select tStAdm from File_BU WHERE T01.TsStat = TsStat)             
WHERE EXISTS (SELECT 1 
		FROM File_BU 
		WHERE TsStat = T01.TsStat
		AND tStADM <> T01.tStADM)

I can't think of a simpler way to do it. SQL server allows the use of JOINs in UPDATEs, and I think DB2 on some platforms has a MERGE statement that could probably simplify this task, but those commands are not part of the SQL standard.