Re: Parsing speed (was Re: pgstats_initstats() cost) - Mailing list pgsql-hackers

From Christopher Kings-Lynne
Subject Re: Parsing speed (was Re: pgstats_initstats() cost)
Date
Msg-id 029501c36559$1a23e2c0$2800a8c0@mars
Whole thread Raw
In response to Re: Parsing speed (was Re: pgstats_initstats() cost)  (Bruce Momjian <pgman@candle.pha.pa.us>)
Responses Re: Parsing speed (was Re: pgstats_initstats() cost)
List pgsql-hackers
> > It might possibly make sense for a libpq routine that exposes Parse
> > to actually do Parse followed by Describe Statement; that would allow
> > it to give back (a) an indication of the number and types of parameters
> > needed by the statement, and (b) an indication of the column set to be
> > returned, if it's a SELECT.  However, the protocol doesn't tell anything
> > about the type of a non-SELECT statement.  In any case, this would
> > require more invention and coding than I care to do at this point in
> > the release cycle (since there's no support in the guts of libpq for
> > accepting ParameterDescription messages from the backend).  If that's
> > what we think we want, we'd better put it on the wish-list for 7.5.

If we had a Parse function, then we at phpPgAdmin could allow Reports to
contain parameters, and detect as such, and then when they run their report,
they can enter the values for that run.

Chris



pgsql-hackers by date:

Previous
From: "Christopher Kings-Lynne"
Date:
Subject: failure in latest cvs
Next
From: Tom Lane
Date:
Subject: Re: failure in latest cvs