Horton

Re: path toward faster partition pruning

Beena Emerson <memissemerson@gmail.com>

View thread View on postgr.es
From: Beena Emerson <memissemerson@gmail.com>
To: Amit Langote <Langote_Amit_f8@lab.ntt.co.jp>
Cc: Rajkumar Raghuwanshi <rajkumar.raghuwanshi@enterprisedb.com>, David Rowley <david.rowley@2ndquadrant.com>, Dilip Kumar <dilipbalaut@gmail.com>, Pg Hackers <pgsql-hackers@postgresql.org>
Date: 2017-10-25T09:18:10Z
Lists: pgsql-hackers 
Hello Amit,

Thanks for the updated patches

On Wed, Oct 25, 2017 at 1:07 PM, Amit Langote
<Langote_Amit_f8@lab.ntt.co.jp> wrote:
> On 2017/10/25 15:47, Amit Langote wrote:
>> On 2017/10/24 1:38, Beena Emerson wrote:
>>> I had noticed this and also that this crash:
>>>
>>> tprt PARTITION BY RANGE(Col1)
>>>        tprt_1 FOR VALUES FROM (1) TO (50001) PARTITION BY RANGE(Col1)
>>>               tprt_11 FOR VALUES FROM (1) TO (10000),
>>>               tprt_1d DEFAULT
>>>        tprt_2 FOR VALUES FROM (50001) TO (100001)
>>>
>>> EXPLAIN (COSTS OFF) SELECT * FROM tprt WHERE col1 BETWEEN 20000 AND 70000;
>>> server closed the connection unexpectedly
>>> This probably means the server terminated abnormally
>>> before or while processing the request.
>>> The connection to the server was lost. Attempting reset: Failed.
>>> !>
>>
>> ...and this (crash) were due to bugs in the 0005 patch.
>
> [ .... ]
>
>> Should be fixed in the attached updated version.
>
> Oops, not quite.  The crash that Beena reported wasn't fixed (or rather
> reintroduced by some unrelated change after once confirming it was fixed).
>
> Really fixed this time.
>

The crashes are fixed. However, handling of DEFAULT partition in
various queries is not proper.

Case 1: In this case default should be selected.
 DROP TABLE tprt;
  CREATE TABLE tprt (col1 int, col2 int) PARTITION BY range(col1);
  CREATE TABLE tprt_1 PARTITION OF tprt FOR VALUES FROM (1) TO (50001)
PARTITION BY list(col1);
  CREATE TABLE tprt_11 PARTITION OF tprt_1 FOR VALUES IN (20000, 25000);
  CREATE TABLE tprt_12 PARTITION OF tprt_1 FOR VALUES IN (50000, 35000);
  CREATE TABLE tprt_13 PARTITION OF tprt_1 FOR VALUES IN (10000);
  CREATE TABLE tprt_1d PARTITION OF tprt_1 DEFAULT;


postgres=#  EXPLAIN (COSTS OFF) SELECT * FROM tprt WHERE col1 < 10000;
        QUERY PLAN
--------------------------
 Result
   One-Time Filter: false
(2 rows)


Case 2: In this case DEFAULT need not be selected.

DROP TABLE  tprt;
  CREATE TABLE tprt (col1 int, col2 int) PARTITION BY range(col1);
  CREATE TABLE tprt_1 PARTITION OF tprt FOR VALUES FROM (1) TO (50001)
PARTITION BY range(col1);
  CREATE TABLE tprt_11 PARTITION OF tprt_1 FOR VALUES FROM (1) TO (10000);
  CREATE TABLE tprt_12 PARTITION OF tprt_1 FOR VALUES FROM (10000) TO (20000);
  CREATE TABLE tprt_13 PARTITION OF tprt_1 FOR VALUES FROM (20000) TO (30000);
  CREATE TABLE tprt_1d PARTITION OF tprt_1 DEFAULT;
  INSERT INTO tprt SELECT generate_series(1,50000), generate_series(1,50000);

postgres=#  EXPLAIN (COSTS OFF) SELECT * FROM tprt WHERE col1 < 10000;
           QUERY PLAN
--------------------------------
 Append
   ->  Seq Scan on tprt_11
         Filter: (col1 < 10000)
   ->  Seq Scan on tprt_1d
         Filter: (col1 < 10000)
(5 rows)


-- 

Beena Emerson

EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Commits

  1. Fix assorted partition pruning bugs

  2. Make gen_partprune_steps static

  3. Remove useless 'default' clause

  4. Reorganize partitioning code

  5. Use custom hash opclass for hash partition pruning

  6. Blindly attempt to fix sepgsql tests broken due to 9fdb675fc5.

  7. Attempt to fix endianess issues in new hash partition test.

  8. Faster partition pruning

  9. For partitionwise join, match on partcollation, not parttypcoll.

  10. Revise API for partition bound search functions.

  11. Revise API for partition_rbound_cmp/partition_rbound_datum_cmp.

  12. Fix possible crash in partition-wise join.

  13. Refactor code for partition bound searching

  14. New C function: bms_add_range

  15. Add extensive tests for partition pruning.

  16. Add null test to partition constraint for default range partitions.

  17. Remove BufFile's isTemp flag.

  18. Make OWNER TO subcommand mention consistent

  19. Fix index matching for operators with mixed collatable/noncollatable inputs.