Horton

Re: Partition-wise join for join between (declaratively) partitioned tables

Robert Haas <robertmhaas@gmail.com>

View thread View on postgr.es
From: Robert Haas <robertmhaas@gmail.com>
To: Ashutosh Bapat <ashutosh.bapat@enterprisedb.com>
Cc: Rafia Sabih <rafia.sabih@enterprisedb.com>, Amit Langote <Langote_Amit_f8@lab.ntt.co.jp>, Rajkumar Raghuwanshi <rajkumar.raghuwanshi@enterprisedb.com>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2017-08-03T01:31:56Z
Lists: pgsql-hackers 
On Mon, Jul 31, 2017 at 9:07 AM, Ashutosh Bapat
<ashutosh.bapat@enterprisedb.com> wrote:
> Forgot the patch set. Here it is.

The commit message for 0005 isn't really accurate given that it
follows 0004.  I think you could just flatten 0005 and 0006 into one
patch.

Reviewing those together:

- Existing code does partdesc = RelationGetPartitionDesc(relation) but
this has got it as part_desc.  Seems better to be consistent.
Likewise existing variables for PartitionKey are key or partkey, not
part_key.

- get_relation_partition_info has a useless trailing return.

- Instead of adding nparts, boundinfo, and part_oids to RelOptInfo,
how about just adding partdesc?  Seems cleaner.

- pkexprs seems like a potentially confusing name, since PK is widely
used to mean "primary key" but here you mean "partition key".  Maybe
partkeyexprs.

- build_simple_rel's matching algorithm is O(n^2).  We may have talked
about this problem before...

- This patch introduces some bits that are not yet used, like
nullable_pkexprs, or even the code to set the partition scheme for
joinrels.  I think perhaps some of that logic should be moved from
0008 to here - e.g. the initial portion of
build_joinrel_partition_info.

There may be more, but I've run out of energy for tonight.

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

Commits

  1. Basic partition-wise join functionality.

  2. Assorted preparatory refactoring for partition-wise join.

  3. Teach adjust_appendrel_attrs(_multilevel) to do multiple translations.

  4. Stamp 10beta2.

  5. Eat XIDs more efficiently in recovery TAP test.

  6. Abstract logic to allow for multiple kinds of child rels.

  7. Implement SortSupport for macaddr data type

  8. Attempt to stabilize grouping sets regression test plans.

  9. Teach xlogreader to follow timeline switches

  10. Don't scan partitioned tables.

  11. Fix grammar.

  12. postgres_fdw: Push down FULL JOINs with restriction clauses.

  13. Some preliminary refactoring towards partitionwise join.

  14. contrib/amcheck needs RecentGlobalXmin to be PGDLLIMPORT'ified.

  15. Print test parameters like "foo: 123", and results like "foo = 123".