>I'd be very interested to get a general view on this behaviour... the fact that 'Workarounds' have to be employed indicates a good reson for defining this as a 'Bug' rather than a 'Design Feature'
>
>regards,
>

Gerard,

I would disagree with you on this. SQL SELECT is supposed to get data from the table on disk, the result set is supposed to reflect the data in the table at the time of the SELECT execusion. Buffered changes have not been committed to the table (they may ultimately be committed or rejected) and therefore cannot be considered as the values in the table.

If the SELECT were to use the buffered data then two stations could run the same SELECT statement at the same time and get different results. I wouldn't call this a bug at all.