>>From what I have read I was under the impression that there wouldn't be any performance increase using INDEXSEEK() over SEEK(). I have a looping counting and totaling procedure where all I did was change a SEEK() to INDEXSEEK() with the move record pointer parameter set to .T., the time the procedure took with the SEEK() was 363 seconds, after changing to INDEXSEEK() the procedure takes 8 seconds. This occurs no matter if I reboot after each run or not, so caching isn't an issue.
>
>Sounds interesting. May be the team never rewrote the seek function internally but the indexseek code is better.
>
>The VFP guys seems to always get another way to increase performance! Look how SQL server performed 5 years ago and now, as a proof! 8-)

The INDEXSEEK() function is much faster than the SEEK() because INDEXSEEK does not move the record pointer, while SEEK does move the record pointer to the first occurance. There is a performance gain to find matching records with INDEXSEEK and if found, then do a SEEK.
The gains are verly obvious over a netwokt, and very impressive over the web.

If you have access to "The Hackers Guide to VFP 6.0" the subject is covered in depth.