Profiler.

By your reaction I'll take that as a no .

GO

SET SHOWPLAN_ALL ON

GO

DECLARE @demo TABLE (ID INT NOT NULL IDENTITY(1,1) PRIMARY KEY CLUSTERED)

INSERT INTO @demo DEFAULT VALUES

INSERT INTO @demo DEFAULT VALUES

INSERT INTO @demo DEFAULT VALUES

INSERT INTO @demo DEFAULT VALUES

INSERT INTO...

Had any traces running when this happened?

BTW, this doesn't work either :

CREATE TABLE #Demo

(

 ID INT NOT NULL IDENTITY (1,1)

)

GO

ALTER TABLE #Demo

ALTER COLUMN ID INT NOT NULL IDENTITY(200,2)

GO

DROP TABLE...

Holly crap, I'd like to see that statement.

4096 colums from 256 tables , 10 M rows each, no index all heaps !

Hey Sergiy, I'm guessing you are using this method for performance.  Is you're math version faster that my varchar version (no time to set up a test

What I was suggesting with a set based aproach was to process all rows with a single update statement.  Or if the operation was too big, then break it into...

I already ported an application from FoxPro (not even sure of the name but it used lists too).  I had to take a few hours to create a lookup table...

are you really talking about the physical file size or the space used in the log file??

A complete query of what Bhudev?

So you're already past 25 year old too Sergiy?

I guess someone had the same unfortunate problem as this one and worked for a few days for digging out the answer .

That happens when the server has no statistics or outdated statistics.  As you did, using a temp table with a clustered index created statistics for that table, hence removing the...

Doesn't help me... maybe someone else may have something.

Good luck with that!

Even an orable DBA could answer this one.

THREAD CLOSED!