I read somewhere that Temp tables - if you take aside the time taken to create them - are far faster than permanent tables for inserts, updates, and deletes. This is because transactions aren't logged on the temp tables so speed increase apparently should be 3 to 4 times faster. Yesterday, I tested this with side by side insert comparisons on a permanent table vs a temporary table - inserting about 10,000 rows. Using Query Analyser and the Execution Plan viewer, the two inserts came out at taking 50% of the total time each. What's going on here? Are temp tables faster or not? Or is there a problem with using the Execution Plan for such kind of analysis? I used Show Execution plan, rather than Estimated Execution plan as I wanted exact actual cost to the db. Cheers guys, Dave.