December 20, 2010 at 11:35 pm
Wildcat (12/20/2010)
How could a NOLOCK hint to the target table on the DELETE or UPDATE statement? Make no sense to me.
Precisely. It's just an oversight in the parser:
-- Error 1065:
-- The NOLOCK and READUNCOMMITTED lock hints are not allowed for
-- target tables of INSERT, UPDATE, DELETE or MERGE statements.
UPDATE dbo.Test WITH (NOLOCK) SET B = 2 WHERE B = 1
-- No error, but no effect either
UPDATE dbo.Test SET B = 2 FROM dbo.Test WITH (NOLOCK) WHERE B = 1;
UPDATE T SET B = 2 FROM dbo.Test T WITH (NOLOCK) WHERE B = 1;
The second two forms do not raise an error, which has mislead people into believing that it somehow behaves differently.
This is a myth. The NOLOCK/READCOMMITTED hint has absolutely no effect here.
SQL Server takes (at least) Update locks when scanning for rows to update, regardless of isolation level.
It's the syntax that is deprecated: the behaviour is unchanged.
December 21, 2010 at 12:29 am
SQLkiwi (12/20/2010)
Wildcat (12/20/2010)
How could a NOLOCK hint to the target table on the DELETE or UPDATE statement? Make no sense to me.Precisely. It's just an oversight in the parser:
-- Error 1065:
-- The NOLOCK and READUNCOMMITTED lock hints are not allowed for
-- target tables of INSERT, UPDATE, DELETE or MERGE statements.
UPDATE dbo.Test WITH (NOLOCK) SET B = 2 WHERE B = 1
-- No error, but no effect either
UPDATE dbo.Test SET B = 2 FROM dbo.Test WITH (NOLOCK) WHERE B = 1;
UPDATE T SET B = 2 FROM dbo.Test T WITH (NOLOCK) WHERE B = 1;
The second two forms do not raise an error, which has mislead people into believing that it somehow behaves differently.
This is a myth. The NOLOCK/READCOMMITTED hint has absolutely no effect here.
SQL Server takes (at least) Update locks when scanning for rows to update, regardless of isolation level.
It's the syntax that is deprecated: the behaviour is unchanged.
Agh. My bad. I didn't read that they were trying to do it on an UPDATE or anything other than a SELECT for that matter.
--Jeff Moden
Change is inevitable... Change for the better is not.
December 21, 2010 at 1:05 am
Jeff Moden (12/21/2010)
Agh. My bad. I didn't read that they were trying to do it on an UPDATE or anything other than a SELECT for that matter.
No worries 🙂
December 22, 2010 at 11:03 am
SQLkiwi (12/20/2010)
Wildcat (12/20/2010)
How could a NOLOCK hint to the target table on the DELETE or UPDATE statement? Make no sense to me.Precisely. It's just an oversight in the parser:
-- Error 1065:
-- The NOLOCK and READUNCOMMITTED lock hints are not allowed for
-- target tables of INSERT, UPDATE, DELETE or MERGE statements.
UPDATE dbo.Test WITH (NOLOCK) SET B = 2 WHERE B = 1
-- No error, but no effect either
UPDATE dbo.Test SET B = 2 FROM dbo.Test WITH (NOLOCK) WHERE B = 1;
UPDATE T SET B = 2 FROM dbo.Test T WITH (NOLOCK) WHERE B = 1;
The second two forms do not raise an error, which has mislead people into believing that it somehow behaves differently.
This is a myth. The NOLOCK/READCOMMITTED hint has absolutely no effect here.
SQL Server takes (at least) Update locks when scanning for rows to update, regardless of isolation level.
It's the syntax that is deprecated: the behaviour is unchanged.
Good catch - I read right over the word target. I only use the FROM clause in a "joined update", where nolock can cause problems (when applied to related tables).
Thanks!
----------------------------------------------------------------------------------
Your lack of planning does not constitute an emergency on my part...unless you're my manager...or a director and above...or a really loud-spoken end-user..All right - what was my emergency again?
Viewing 4 posts - 16 through 18 (of 18 total)
You must be logged in to reply to this topic. Login to reply