Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commitfd48a86

Browse files
Doc: clarify the conditions of usable indexes for REPLICA IDENTITY FULL tables.
Commit89e46da allowed REPLICA IDENTITY FULL tables to use an indexon the subscriber during apply of update/delete. This commit clarifiesin the documentation that the leftmost field of candidate indexes mustbe a column (not an expression) that references the published relationcolumn.The source code comments are also updated accordingly.Reviewed-by: Peter Smith, Amit KapilaDiscussion:https://postgr.es/m/CAD21AoDJjffEvUFKXT27Q5U8-UU9JHv4rrJ9Ke8Zkc5UPWHLvA@mail.gmail.comBackpatch-through: 16
1 parent0fef877 commitfd48a86

File tree

3 files changed

+13
-16
lines changed

3 files changed

+13
-16
lines changed

‎doc/src/sgml/logical-replication.sgml

Lines changed: 6 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -134,12 +134,12 @@
134134
to replica identity <literal>FULL</literal>, which means the entire row becomes
135135
the key. When replica identity <literal>FULL</literal> is specified,
136136
indexes can be used on the subscriber side for searching the rows. Candidate
137-
indexes must be btree, non-partial, andhave at least one column reference
138-
(i.e. cannot consist of only expressions). These restrictions
139-
on the non-unique index properties adhere to some of the restrictions that
140-
are enforced for primary keys. If there are no such suitable indexes,
141-
the search on the subscriber side can be very inefficient, therefore
142-
replica identity <literal>FULL</literal> should only be used as a
137+
indexes must be btree, non-partial, andthe leftmost index field must be a
138+
column (not an expression) that references the published table column. These
139+
restrictionson the non-unique index properties adhere to some of the
140+
restrictions thatare enforced for primary keys. If there are no such
141+
suitable indexes,the search on the subscriber side can be very inefficient,
142+
thereforereplica identity <literal>FULL</literal> should only be used as a
143143
fallback if no other solution is possible. If a replica identity other
144144
than <literal>FULL</literal> is set on the publisher side, a replica identity
145145
comprising the same or fewer columns must also be set on the subscriber

‎src/backend/executor/execReplication.c

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -47,9 +47,9 @@ static bool tuples_equal(TupleTableSlot *slot1, TupleTableSlot *slot2,
4747
*
4848
* Returns how many columns to use for the index scan.
4949
*
50-
* This is not generic routine,it expects theidxrelto bea btree, non-partial
51-
*and have at least one column reference (i.e. cannot consist of only
52-
*expressions).
50+
* This is not generic routine, idxrelmust bePK, RI, or an index that can be
51+
*used for REPLICA IDENTITY FULL table. See FindUsableIndexForReplicaIdentityFull()
52+
*for details.
5353
*
5454
* By definition, replication identity of a rel meets all limitations associated
5555
* with that. Note that any other index could also meet these limitations.

‎src/backend/replication/logical/relation.c

Lines changed: 4 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -779,9 +779,10 @@ RemoteRelContainsLeftMostColumnOnIdx(IndexInfo *indexInfo, AttrMap *attrmap)
779779

780780
/*
781781
* Returns the oid of an index that can be used by the apply worker to scan
782-
* the relation. The index must be btree, non-partial, and have at least
783-
* one column reference (i.e. cannot consist of only expressions). These
784-
* limitations help to keep the index scan similar to PK/RI index scans.
782+
* the relation. The index must be btree, non-partial, and the leftmost
783+
* field must be a column (not an expression) that references the remote
784+
* relation column. These limitations help to keep the index scan similar
785+
* to PK/RI index scans.
785786
*
786787
* Note that the limitations of index scans for replica identity full only
787788
* adheres to a subset of the limitations of PK/RI. For example, we support
@@ -796,10 +797,6 @@ RemoteRelContainsLeftMostColumnOnIdx(IndexInfo *indexInfo, AttrMap *attrmap)
796797
* none of the tuples satisfy the expression for the index scan, we fall-back
797798
* to sequential execution, which might not be a good idea in some cases.
798799
*
799-
* We also skip indexes if the remote relation does not contain the leftmost
800-
* column of the index. This is because in most such cases sequential scan is
801-
* favorable over index scan.
802-
*
803800
* We expect to call this function when REPLICA IDENTITY FULL is defined for
804801
* the remote relation.
805802
*

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp