>You can pull down all the possibly related SQL data into another VFP cursor, then do the JOIN on the two cursors. Or you can push your cursor into a SQL table and then do the JOIN in SQL and return the results. Or you can put the cursor into a VFP table and then have a SQL SP do the JOIN using OPENROWSET to access the VFP data directly over ODBC. There may even be a few more ways to do it, but the way I'd recommend (without know how much data is involved -- which could alter the option chosen) is the first one. Pull down the SQL data and JOIN the two cursors in VFP. The other options are less reliable or more klunky IMO.

The first possibilty is a no go because we are talking about millions of records.

The second possiblity, I would like to avoid creating the cursor INTO TABLE instead in order to prepare it for the SQLExec() INNER JOIN clause as it will create a file on disk.

Basically, I would like to find a way to handle the cursor with a SQLExec() approach without having to create a file on disk.