>>The changes aren't written to the SQL on newly inserted records, but are on the records which existed.
>>
>>I'll eventually catch the time to revisit this and check the error codes and whether the adapter tried to insert the same record twice or what.
>
>Table has a first buffer record level that it is flushed with a record move.
>With a go to recno() you can force this flush into the table buffer.

Back in the day we used a skip 0 for the same effect. Not sure it would change the record's status in the case of a SPT cursor with cursoradapter, but then why not. I was thinking of that but then thought I won't need it, this job is done already.

But, of course, the customer forgot to tell me they wanted yet another column they didn't provide the first time, so, well, I'll soon check the theories :).