Re: [SQL] More efficient DELETE ... ?

> The Hermit Hacker <scrappy(at)hub(dot)org> writes:
> > This looks painful ...
>
> > DELETE FROM webhit_referer_raw
> > WHERE oid IN ( SELECT w.oid
> > FROM webhit_referer_raw w, referrer_data r
> > WHERE stat_date < 'Jan 17 2000 15:05:00'
> > AND w.referrer_url = r.referrer );
>
> I believe you can do
>
> DELETE FROM webhit_referer_raw
> WHERE referrer_url = referrer_data.referrer
> AND referrer_data.stat_date < 'Jan 17 2000 15:05:00';
>
> (I assume stat_date is in referrer_data, otherwise this'd be easy.)
>
> This is an example where Postgres' willingness to create implicit FROM
> clause entries is a win ... AFAIK you couldn't do it in standard SQL,
> since for some reason DELETE doesn't take a FROM clause.

Good point. I had forgotten all about that.

Should I cover that in my book?

What about the NOTICE we were going to add when this happens. You will
see a WARN_FROM define in parser/*.c. I had to disable the message
because the regression tests were failing. They have to be regenerated
after I do this, but I can't do that here. Do people still want the
message? Can someone regenerate the regression output after I am done?

In fact, someone can just remove the #ifdef WARN_FROM defines around the
code and regenerate the regression tests and all will stay in sync.

--
Bruce Momjian | http://www.op.net/~candle
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026