I am seeing something strange. When I do a count(*) from table where column IS NULL, my result is every row in the table. This is a 46 million row table and I know that not every row contains a null in the the specific column that I am testing for NULL. I know that there are about 2 million that do. If I do a select top 3000000 * into #temp where column is NULL, I get my 2 million records inserted into the temp table. But why will count(*) not show me a count of 2 million? Does SQL Server have problem with large tables in this regard? Thanks!