I have never used this syntax, but I would also add the question " what happens in a multi user environment"? If it always looks at just the last row, then it would not always be the intended user's inserted row, but perhaps another process's.
I once heard that it was usually best to avoid using the record position/row count in a table when possible if adhering to SQL structure. Buy hey, I still use the COUNT = LAST in a few applications!
-Bob
-------------- Original message --------------
>
> Perhaps someone can offer some insight.
>
> I've always assumed that COUNT = INSERT (in a WHERE clause) is only valid
> immediately after the user performs an insert.
>
> The R:Base documentation states that COUNT = INSERT will return the last
> row inserted by the user.
>
> Question #1 is "How long is that valid?" That is, a user inserts a row
> into a table. Then the user goes and does some other stuff, none of which
> involves inserts or updates to other tables, and the code then checks for
> COUNT = INSERT on the table where the row was created. What will cause
> COUNT = INSERT to become no longer valid in this context?
>
> Question #2 is "What happens to COUNT = INSERT when the user inserts a row
> into table A, inserts a row into table B, then checks the COUNT = INSERT on
> table A?"
>
> Question #3 is "What should COUNT = INSERT return if the user HAS NOT
> inserted a row?" (It appears to return the last row in the table. Should
> this be valid?)
>
>
>
>
>
> Emmitt Dove
> Manager, DairyPak Business Systems
> Blue Ridge Paper Products, Inc.
> 40 Lindeman Drive
> Trumbull, CT 06611
> (203) 673-2231
> [EMAIL PROTECTED]
> [EMAIL PROTECTED]
>
