Jaime-

>What i meant is either you have the table/cursor in the DE, or else you should issue a USE... to the tables involved in the SELECT then run the select. I know you could run the SELECT without using the tables, but then you should either enter the full path of the tables involved inside the SELECT statement, or your environment should have a PATH statement to the data directory where the tables reside. Since i didn't know for sure if Nir did set up the environment this way, i rather recommended to USE the tables before...

I disagree. It's better, IMO, to just issue the select. Select pulls from the tables on disk, anyway. And, for folks who end up with double buffering (buffed view and buffered tables) it can be very confusing for them and hard to debug why tableupdates don't seem to work.

Naturally, Fox has to be able to find the tables. That would come from knowing how you set up your pathing and whether database containers are open.