Horton

Re: Declarative partitioning - another take

Amit Langote <langote_amit_f8@lab.ntt.co.jp>

View thread View on postgr.es
From: Amit Langote <Langote_Amit_f8@lab.ntt.co.jp>
To: Robert Haas <robertmhaas@gmail.com>
Cc: Dmitry Ivanov <d.ivanov@postgrespro.ru>, Michael Paquier <michael.paquier@gmail.com>, Andres Freund <andres@anarazel.de>, Erik Rijkers <er@xs4all.nl>, Amit Langote <amitlangote09@gmail.com>, Rajkumar Raghuwanshi <rajkumar.raghuwanshi@enterprisedb.com>, Ashutosh Bapat <ashutosh.bapat@enterprisedb.com>, Pg Hackers <pgsql-hackers@postgresql.org>, pgsql-hackers-owner@postgresql.org
Date: 2016-12-16T08:42:54Z
Lists: pgsql-hackers

Attachments

On 2016/12/16 17:02, Amit Langote wrote:
> [PATCH 2/7] Change how RelationGetPartitionQual() and related code works
> 
> Since we always want to recurse, ie, include the parent's partition
> constraint (if any), get rid of the argument recurse.
> 
> Refactor out the code doing the mapping of attnos of Vars in partition
> constraint expressions (parent attnos -> child attnos).  Move it to a
> separate function map_partition_varattnos() and call it from appropriate
> places.  It previously used to be done in get_qual_from_partbound(),
> which would lead to wrong results in certain multi-level partitioning
> cases, as the mapping would be done for immediate parent-partition pairs.
> Now in generate_partition_qual() which is the workhorse of
> RelationGetPartitionQual(), we first generate the full expression
> (considering all levels of partitioning) and then do the mapping from the
> root parent to a leaf partition.  It is also possible to generate
> partition constraint up to certain non-leaf level and then apply the
> same to leaf partitions of that sub-tree after suitable substitution
> of varattnos using the new map_partition_varattnos() directly.
> 
> Bug fix: ATExecAttachPartition() failed to do the mapping when attaching
> a partitioned table as partition. It is possible for the partitions of
> such table to have different attributes from the table being attached
> and/or the target partitioned table.

Oops, PATCH 2/7 attached with the previous email had a bug in it, whereby
map_partition_varattnos() was not applied to the partition constraint
expressions returned directly from the relcache (rd_partcheck) copy.
Attaching just the updated (only) PATCH 2 which fixes that.  Patches
1,3,4,5,6,7/7 from the previous email [1] are fine.  Sorry about that.

Thanks,
Amit

[1]
https://www.postgresql.org/message-id/c820c0eb-6935-6f84-8c6a-785fdff130c1@lab.ntt.co.jp

Commits

  1. Fix typo.

  2. Document trigger-firing behavior for inheritance/partitioning.

  3. Fire per-statement triggers on partitioned tables.

  4. Set ecxt_scantuple correctly for tuple routing.

  5. Fix interaction of partitioned tables with BulkInsertState.

  6. Avoid core dump for empty prepared statement in an aborted transaction.

  7. Fix some problems in check_new_partition_bound().

  8. Remove unnecessary arguments from partitioning functions.

  9. Fix reporting of constraint violations for table partitioning.

  10. Fix tuple routing in cases where tuple descriptors don't match.

  11. Invalid parent's relcache after CREATE TABLE .. PARTITION OF.

  12. Doc: improve documentation about inheritance.