Re: Why ContinueUpdateOnError is not implemented in npgsql

[Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>]

"Andrus" <kobruleht2(at)hot(dot)ee> writes:
> ContinueUpdateOnError requires that transaction continues after error. In 
> this case .NET can mark all bad rows in DataGrid. User can see all errors 
> together.
> ...
> There are only two ways fix this issue.
> NpgsqlDataAdapter must invoke automatic ROLLBACK after each error
> or use checkpoints before each command.

Yup, a savepoint before each command is required if that's the behavior
you want.  Yes, that adds overhead.  The reason it's not automatic is
exactly that it adds overhead, which many applications don't need or
want.  But if you have to have it, that's what you do.

			regards, tom lane