Horton

Re: [HACKERS] Runtime Partition Pruning

David Rowley <david.rowley@2ndquadrant.com>

View thread View on postgr.es
From: David Rowley <david.rowley@2ndquadrant.com>
To: Alvaro Herrera <alvherre@alvh.no-ip.org>
Cc: Beena Emerson <memissemerson@gmail.com>, Amit Langote <Langote_Amit_f8@lab.ntt.co.jp>, Robert Haas <robertmhaas@gmail.com>, amul sul <sulamul@gmail.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>, Dilip Kumar <dilipbalaut@gmail.com>
Date: 2018-01-04T23:58:24Z
Lists: pgsql-hackers 
On 5 January 2018 at 05:37, Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
> I tried this patch (applying it on Amit's last current version on top of
> 4e2970f8807f which is the latest it applies to) and regression tests
> fail with the attached diff; in all cases it appears to be an off-by-one
> in row count.  Would you please give it a look?

Looks like it's down to ExplainPropertyFloat() having
machine-dependent behaviour.

On the machine that I was working with when testing this the following
code outputs "1"

#include <stdio.h>

double nfiltered = 1.0;
double nloops = 2.0;

int main(void)
{
printf("%.*f", 0, nfiltered / nloops);
return 0;
}

but on your machine it must be outputting "0"?

I'm not sure if I should just try to have even row numbers in the
tests, or if we should be fixing EXPLAIN so it is consistent.

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

Commits

  1. Initialize ExprStates once in run-time partition pruning

  2. Add bms_prev_member function

  3. Support partition pruning at execution time

  4. Document partprune.c a little better

  5. Fix crash in pg_replication_slot_advance

  6. Add parallel-aware hash joins.

  7. Fix typo.

  8. Allow --with-bonjour to work with non-macOS implementations of Bonjour.

  9. Mention need for --no-inc-recursive in rsync command