Hi Thom,

You could grab the cursor from the back-end with the sqlexec(), and then copy it into a table within the user's work area for the user's exclusive use. Then the user could add new records and update existing records in the local table. When done, the user could send the inserts and updates back to the back-end from the local table. However, this solution might not be as good as the one your currently using.

I have written a VFP Front-End client to a postgres back-end server that uses this technique. It is an example application complete with source code and it is free. If your interested you might take a look. I would be interested in any suggestion you might have on improving my technique. The example app can be found at:

http://www.smvfp.com

When you enter my site, look for option menu 19.

I have been interested in parameterized view and got around, just this morning, to including one example of using a parameterized view along with my other two examples of using SQL pass throughs in my example app.

Hi Thom,

It me again.

By the way, is there a potential problem in using remove view when working with a non-native back-end server like MSSQL or Postgres? The fact that you feel a need to be unlinked from the back-end server worries me a little. Is it possible that you are unlinked from the back-end, but because VFP sets the buffering to optomistic 3, VFP may be issuing a SQLexec() everytime you advance to a new row. I think a view is exactly like SQL pass through, except VFP does a lot of the work for you in the background like identifying records to be updated and inserted, prasing the commands and sending the command to the back-end server.