On Sat, 26 Apr 2003, Bruce Momjian wrote:
> Rajesh Kumar Mallah wrote:
> > On Saturday 26 Apr 2003 8:35 am, Tom Lane wrote:
> > > Jan Wieck <JanWieck@Yahoo.com> writes:
> > > > The question back is from where does such a query with a comment come?
> > > > If it is a standard client like psql, it could happen someday that this
> > > > client does some filtering.
> >
> >
> > Yes its thru psql only , i run a SQL batch using
> > psql -e -h db -f batch.sql
> >
> > In case future version of psql strips them it would
> > still be nice if libpq continue supporting it
> > i feel it would be a nice feature for keeping track of
> > queries.
>
> [ After not replying to the proper question before, I will try again. :-) ]
>
Thanks Bruce,
But if i were to use DBD::Perl (which probably uses libpq)
is it safe to pass query indentifier in comments which u
said are not stripped by psql eg
/* AppName: Xyzt */ Select foo from bar where cw='t' ;
for years to come ?
why i ask is that i was always curious of the slow
long running queries in pg_stat_activity and not able
to figure out the source had frustrated me , now
this method if it works seems to me as a viable solution.
regds
mallah.
> The actual stripping of comments in psql is a little more complex. For
> example, this will not strip:
>
> SELECT /* test */ 1;
>
> while this will strip comments:
>
> SELECT 1; /* test */
>
> and leading comments are not stripped either:
>
> /* test */ SELECT 1;
>
> so if the comment is _inside_ a query, it will be OK, but outside and
> trailing, psql processes the comment itself so it can display the proper
> prompting:
>
> test=> /*
> test*> test
>
> ^
> test*> */
>
>
