>Mike,
>
>Try keymatch() instead. That is definitely faster than seek.
>
>Walter,
>
>>Hey all
>>
>>Just FYI
>>
>>I have rarely had a need to use indexseek. I read it's "fast". I thought it's faster than other things, but it's not faster than seek. The only benefit I can see in it is it does a seek without moving the record pointer. That's good for uniqueness checking during adding/editing of a record.
>>
>>On a 27 million record table, indexseek and seek are almost exactly the same speed.

It's fast alright. It moves the record pointer and puts it back, which might give me a bit of grief. I'll check. Also to get the index #, I'd have to use tagno(). That would slow the process down.