We've seen this happen a couple times on joins to derived tables, but never seen it happen from two permanent tables with an inner join. This code looks like it was transplanted from MS Access. Is that correct? If so, sometimes there can be errors in the transfer from Jet to T-SQL. It may be worth rewriting the code in a slightly more readable JOIN syntax. You don't need all the NOLOCK statements, or the LEFT LOOP JOIN (which looks suspicious already because you are INNER JOINING the same 2 tables in the EXISTS...). If possible, could you post either a SHOWPLAN or a the table schemas (or both) so we can take a deeper look at this? Thanks, Jay.