From: | Andres Freund <andres(at)anarazel(dot)de> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | pgsql-committers(at)postgresql(dot)org |
Subject: | Re: pgsql: Avoid use of float arithmetic in bipartite_match.c. |
Date: | 2015-08-23 17:44:37 |
Message-ID: | 20150823174437.GJ8552@awork2.anarazel.de |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-committers |
On 2015-08-23 13:30:45 -0400, Tom Lane wrote:
> Andres Freund <andres(at)anarazel(dot)de> writes:
> > Not sure why you replaced n by k?
>
> I thought it was possible to confuse it with the "n"'s used in the
> previous line to denote the graph sizes.
>
> > the nodes are 1..n, so the adjacency list should be as well (or the
> > other way round).
>
> No, I meant them to be different. Do you think the other way is better?
Well, using the same index seemed to comfortably explain that the size
of the individual adjacency mappings depends on the number of nodes in
V. But for that to really make sense there'd need to be a different
indexes for U/V, which I now see wasn't the case case before
either. As currently used they'll be the same anyway...
- Andres
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2015-08-23 19:15:52 | pgsql: Reduce number of bytes examined by convert_one_string_to_scalar( |
Previous Message | Tom Lane | 2015-08-23 17:30:45 | Re: pgsql: Avoid use of float arithmetic in bipartite_match.c. |