>>I use locate for field = val on tables with 100,000 to over 1 million records all the time and I can't see any difference between seek/locate when optimized. The nice thing about locate is the index is implied, you do have to set/reset the order. Locate work best when NO ORDER is set.
>
>I've just tested on a large table, and you're right, it seems as fast as SEEK now. There's a note in Hacker's that it used to be slow, so that's what I'm referring to, I guess. Also, looking over some old code that others did here that's slow, I see that often the field was placed on the right side of the '=', and that won't optimize, clearly (back then it didn't matter, though).
>
>But LOCATE does seem plenty fast now, anyway...I still prefer the SEEK() or INDEXSEEK() (in vfp6), since you have more flexibility with the extra parameters, however...no SELECT, no ORDER.
I had never heard/used indexseek, what other language give so many viable options.
-TomC