>>Only because it's a lot faster to index 0 items than to index 10,000 items.
>>
>>Alan
>
>Interesting...
>I've read that when adding a large number of records to a table its faster to remove indexes, add records and then to recreate the indexes. Isn't it the same in this situation?
>Kamil.

That's what Mike is saying. I have to admit I've nevere actually timed it. From having done it both ways in various programs, It's always seemed to me to be faster doing it the say I said. Mike's code shows it to be the other way around. The actual timing shows that it's faster to load the data first and then index. It's weird because, not only has it seemed faster to me to do it the other way, but it's seemed to be a LOT faster.

Time to change some code.

Alan