Re: Faster methods for getting SPI results (460% improvement)

On 24 January 2017 at 11:23, Jim Nasby <Jim(dot)Nasby(at)bluetreble(dot)com> wrote:

> I finally got all the kinks worked out and did some testing with python 3.
> Performance for my test [1] improved ~460% when returning a dict of lists
> (as opposed to the current list of dicts). Based on previous testing, I
> expect that using this method to return a list of dicts will be about 8%
> slower. The inconsistency in results on 2.7 has to do with how python 2
> handles ints.

Impressive results.

> I think the last step here is to figure out how to support switching between
> the current behavior and the "columnar" behavior of a dict of lists.

That sounds like it'd be much better approached as a separate, later patch.

If I understand you correctly, you propose to return the resultset

a b
1 10
2 20

which is currently returned as

[ {"a":1, "b":10}, {"a":2, "b":20} ]

instead as

{ "a": [1, 2], "b": [10, 20] }

?

If so I see that as a lot more of a niche thing. I can see why it'd be
useful and would help performance, but it seems much more disruptive.
It requires users to discover it exists, actively adopt a different
style of ingesting data, etc. For a 10%-ish gain in a PL.

I strongly suggest making this design effort a separate thread, and
focusing on the SPI improvements that give "free" no-user-action
performance boosts here.

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