Re: "RETURNING PRIMARY KEY" syntax extension

On 14/06/12 20:58, Jochem van Dieten wrote:
> On Thu, Jun 12, 2014 at 12:25 PM, Ian Barwick wrote:
>
> On 14/06/12 18:46, Jochem van Dieten wrote:
> > I haven't checked the code, but I am hoping it will help with the problem
> > where a RETURNING * is added to a statement that is not an insert or update
> > by the JDBC driver. That has been reported on the JDBC list at least twice,
> > and the proposed workaround is neither very elegant nor very robust:
> > https://groups.google.com/forum/#!msg/pgsql.interfaces.jdbc/7WY60JX3qyo/-v1fqDqLQKwJ
>
> Unfortunately that seems to be a JDBC-specific issue, which is outside
> of the scope of this particular patch (which proposes additional server-side
> syntax intended to make RETURNING * operations more efficient for
> certain use cases, but which is in itself not a JDBC change).
>
>
> But the obvious way to fix the JDBC issue is not to fix it by adding a 'mini parser' on
> the JDBC side, but to make SELECT ... RETURNING PRIMARY KEY a regular select that silently
> ignores the returning clause and doesn't throw an error on the server-side.
>
> That might still be outside the scope of this particular patch, but it would provide
> (additional) justification if it were supported.

That would be adding superfluous, unused and unusable syntax of no potential value
(there is no SELECT ... RETURNING and it wouldn't make any sense if there was) as a
workaround for a driver issue - not going to happen.

Regards

Ian Barwick

--
Ian Barwick http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services