Re: [HACKERS] Performance improvement for joins where outer side is unique

On 26 January 2017 at 04:56, Antonin Houska <ah@cybertec.at> wrote:
> I suspect that "inner" and "outer" relation / tuple are sometimes confused in
> comments:
>
>
> * analyzejoins.c:70
>
>  "searches for subsequent matching outer tuples."
>
>
> * analyzejoins.c:972
>
> /*
>  * innerrel_is_unique
>  *        Check for proofs which prove that 'innerrel' can, at most, match a
>  *        single tuple in 'outerrel' based on the join condition in
>  *        'restrictlist'.
>  */
>
>
> * relation.h:1831
>
>         bool            inner_unique;   /* inner side of join matches no more than one
>                                                                  * outer side tuple */

Thanks for looking over the patch. I believe I've fixed all those now.
I'm not too surprised I got some wrong, after all, I did mess up the
subject of this email too! Which did cause quite a bit of confusion.


-- David Rowley                   http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training & Services