>FWIW - when doing complex queries against VFP data I find it MUCH faster to prequery the main tables on simple rushmore optimizable expressions and then use those cursors in my "real" query
>
>select fieldnames from maintable1 where somenum=x into cursor c_maintable1 readwrite nofilter
>
>select fieldnames from maintable2 where somenum=x into cursor c_maintable2 readwrite nofilter
>
>
>select c_maintable1.fieldnames,c_maintable2.fieldnames
>from c_maintable1 inner/left join on c_maintable2 on somecriteria blah blah blah
>
>
>Using this technique I have sped up queries by a factor of 10 or more.

Specially when the maintable1 was big and there were more joins, the factor was about 20... which made all the difference between "slow like molasses" and "immediately"... 12 years ago on a 200MHz machine. Still works nowadays, even on SQL. Sometimes I just get the selection from the big table, and relevant records from another, then join those cursors in Fox, or if I don't really need to show the joined fields, just check them for some records (in a batch process), I scan the first cursor while seek()ing the second. The repeated values aren't traveling across the wire, and I actually don't need them in all records of the first table.

This kind of scenario works every once in a while, and I'm glad it can be done this way - while developing and testing the thing, I have to run it dozens of times a day, so speed gains are important - it's MY time we're talking about :).