Thomas Lockhart writes:
> > > Nothing serious, but I would like to apply a patch to allow IDENT
> > > strings (e.g. 'hour') to be accepted by the SQL92 EXTRACT() function. We
> > > accept those for date_part(), which is what EXTRACT() is translated to
> > > by the parser, and it seems to be a reasonable to the standard.
> > But why does that need to go into 7.1.1?
>
> Does not "need to". But it is non-invasive, extremely low risk, gets the
> behavior to match the docs, and gets it off my desk and into the main
> tree.
Hehe, match the docs? The docs used to be perfectly accurate until you
changed them.
--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter
