Re: 25P02, current transaction is aborted, commands ignored

Well,

The concept of an atomic transaction means that it must either
succeed completely or fail completely. PostgreSQL does this.

Dave
On 2-Apr-06, at 8:08 AM, Amaresh Wakkar wrote:

> Thanks!!
>
> The code works fine after I added savepoints around "Okay,even if
> fails" statement.
>
> What is the idea behind taking this route(i.e all statements
> ignored till end of block) though? Is there any archive thread
> discussing this? In this particular case, the autocommit is off and
> the statement has failed so there is no risk of changes being made
> permanent(presumable PG would rollback at database level the
> changed buffers etc.,) unless the commit() method is invoked. Why
> then explicit rollback is needed? Is it not best if it is left to
> programmer to handle this in exception handling code?
>
> I just had to add two lines of savepoints but I added them for
> making my code work rather than using them for some application
> logic purpose.
>
> Cheers!!
>
> Amaresh Wakkar
>
>
>
> "Oliver Jowett" wrote:
>
>
> babu_moshay wrote:
>
> > In my opinion, if there are reasons to throw 25P02 and abort
> transaction unilaterally, then there are also good reasons not to
> abort it and let programmer take the decision. A switching
> mechanism would have been ideal.
>
> Create a savepoint before the possibly-failing query. If the query
> fails
> in the way you were expecting, roll back to the savepoint and
> continue.
>
> -O
>
> ---------------------------(end of
> broadcast)---------------------------
> TIP 3: Have you checked our extensive FAQ?
>
> http://www.postgresql.org/docs/faq
>
> Indiatimes Email now powered by APIC Advantage. Help!
> Help
>