josh 92884 (10/21/2013)
First post here so hope I don't come of as too newb with this.I have a table on a MSSQL 2000 with about 4700 rows.
My query:
SELECT * FROM Product with (nolock);
only comes back with about 250 rows, sometimes slightly less or more. This query also takes awhile to complete. Longer than these:
SELECT colname FROM Product with (nolock);
SELECT * FROM Product;
which come back with the correct number of records.
I'm a little mystified. What is it about the combination of * and NOLOCK that causes this?
Nothing unusual stands out to me about the table design. It is a clustered index with a few non-clustered indexes on non-key columns. DBCC SHOWCONTIG(Product) looks reasonable.
Anyone have any ideas? or next steps to investigate?
I've never seen WITH(NOLOCK) cause such a large disparity in the number of rows like that before. Any clues in the execution plan?
--Jeff Moden
Change is inevitable... Change for the better is not.