On 8/14/05, Bruce Momjian <pgman@candle.pha.pa.us> wrote:
> Brendan Jurd wrote:
> > > We already have a TODO for this:
> > >
> > > * Add transaction_timestamp(), statement_timestamp(), clock_timestamp()
> > > functionality
> >
> > I like the idea of having a function for statement start time. I
> > think I'll incorporate it into my patch.
> >
Regarding the statement_timestamp() ... if the entire query path is
parser -> rewriter -> planner/optimiser -> executor, what point in
that path would be considered the true start of the "statement"? I
would assume that it's right at the beginning, when it is first handed
to the parser, but that's a very naive assumption.