Re: bug in join? - Mailing list pgsql-hackers

From Tom Lane
Subject Re: bug in join?
Date
Msg-id 7726.1010010596@sss.pgh.pa.us
Whole thread Raw
In response to Re: bug in join?  (Laurette Cisneros <laurette@nextbus.com>)
Responses Re: bug in join?  (Laurette Cisneros <laurette@nextbus.com>)
Re: bug in join?  (Laurette Cisneros <laurette@nextbus.com>)
List pgsql-hackers
Laurette Cisneros <laurette@nextbus.com> writes:
> I will keep trying to recreate it for you.

You could just try
analyze b;analyze d;explain select * from b,d where b.address = d.address;

and repeat until you see the error from EXPLAIN.  Since ANALYZE takes
a random sampling these days, successive loops will in fact produce
slightly different results, and you may be able to recreate the
erroneous state eventually.

The math in eqjoinsel() is not entirely trivial, but I thought I had
convinced myself it was okay.  I need to see a failing example to 
figure out what's wrong with it.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: Bulkloading using COPY - ignore duplicates?
Next
From: Laurette Cisneros
Date:
Subject: Re: bug in join?