database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-07-12 21:01:52 +03:00
Code Activity

Avoid full scan of GIN indexes when possible

Browse Source 
The strategy of GIN index scan is driven by opclass-specific extract_query
method.  This method that needed search mode is GIN_SEARCH_MODE_ALL.  This
mode means that matching tuple may contain none of extracted entries.  Simple
example is '!term' tsquery, which doesn't need any term to exist in matching
tsvector.

In order to handle such scan key GIN calculates virtual entry, which contains
all TIDs of all entries of attribute.  In fact this is full scan of index
attribute.  And typically this is very slow, but allows to handle some queries
correctly in GIN.  However, current algorithm calculate such virtual entry for
each GIN_SEARCH_MODE_ALL scan key even if they are multiple for the same
attribute.  This is clearly not optimal.

This commit improves the situation by introduction of "exclude only" scan keys.
Such scan keys are not capable to return set of matching TIDs.  Instead, they
are capable only to filter TIDs produced by normal scan keys.  Therefore,
each attribute should contain at least one normal scan key, while rest of them
may be "exclude only" if search mode is GIN_SEARCH_MODE_ALL.

The same optimization might be applied to the whole scan, not per-attribute.
But that leads to NULL values elimination problem.  There is trade-off between
multiple possible ways to do this.  We probably want to do this later using
some cost-based decision algorithm.

Discussion: https://postgr.es/m/CAOBaU_YGP5-BEt5Cc0%3DzMve92vocPzD%2BXiZgiZs1kjY0cj%3DXBg%40mail.gmail.com
Author: Nikita Glukhov, Alexander Korotkov, Tom Lane, Julien Rouhaud
Reviewed-by: Julien Rouhaud, Tomas Vondra, Tom Lane
This commit is contained in:
Alexander Korotkov
2020-01-18 01:11:39 +03:00
parent 41c6f9db25
commit 4b754d6c16
10 changed files with 579 additions and 77 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
src/include/access/gin_private.h
View File

@ -303,6 +303,20 @@ typedef struct GinScanKeyData
int32 searchMode;
OffsetNumber attnum;
/*
* An excludeOnly scan key is not able to enumerate all matching tuples.
* That is, to be semantically correct on its own, it would need to have a
* GIN_CAT_EMPTY_QUERY scanEntry, but it doesn't. Such a key can still be
* used to filter tuples returned by other scan keys, so we will get the
* right answers as long as there's at least one non-excludeOnly scan key
* for each index attribute considered by the search. For efficiency
* reasons we don't want to have unnecessary GIN_CAT_EMPTY_QUERY entries,
* so we will convert an excludeOnly scan key to non-excludeOnly (by
* adding a GIN_CAT_EMPTY_QUERY scanEntry) only if there are no other
* non-excludeOnly scan keys.
*/
bool excludeOnly;
/*
* Match status data. curItem is the TID most recently tested (could be a
* lossy-page pointer). curItemMatches is true if it passes the