postgresql/src/backend/statistics/extended_stats.c

516 lines
13 KiB
C
Raw Normal View History

Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
/*-------------------------------------------------------------------------
*
* extended_stats.c
* POSTGRES extended statistics
*
* Generic code supporting statistics objects created via CREATE STATISTICS.
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
*
*
* Portions Copyright (c) 1996-2018, PostgreSQL Global Development Group
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
* Portions Copyright (c) 1994, Regents of the University of California
*
* IDENTIFICATION
* src/backend/statistics/extended_stats.c
*
*-------------------------------------------------------------------------
*/
#include "postgres.h"
#include "access/genam.h"
#include "access/heapam.h"
#include "access/htup_details.h"
#include "catalog/indexing.h"
#include "catalog/pg_collation.h"
#include "catalog/pg_statistic_ext.h"
#include "nodes/relation.h"
#include "postmaster/autovacuum.h"
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
#include "statistics/extended_stats_internal.h"
#include "statistics/statistics.h"
#include "utils/builtins.h"
#include "utils/fmgroids.h"
#include "utils/lsyscache.h"
#include "utils/memutils.h"
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
#include "utils/rel.h"
#include "utils/syscache.h"
/*
* Used internally to refer to an individual statistics object, i.e.,
* a pg_statistic_ext entry.
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
*/
typedef struct StatExtEntry
{
Oid statOid; /* OID of pg_statistic_ext entry */
char *schema; /* statistics object's schema */
char *name; /* statistics object's name */
Bitmapset *columns; /* attribute numbers covered by the object */
List *types; /* 'char' list of enabled statistic kinds */
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
} StatExtEntry;
static List *fetch_statentries_for_relation(Relation pg_statext, Oid relid);
static VacAttrStats **lookup_var_attr_stats(Relation rel, Bitmapset *attrs,
int nvacatts, VacAttrStats **vacatts);
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
static void statext_store(Relation pg_stext, Oid relid,
MVNDistinct *ndistinct, MVDependencies *dependencies,
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
VacAttrStats **stats);
/*
* Compute requested extended stats, using the rows sampled for the plain
* (single-column) stats.
*
* This fetches a list of stats types from pg_statistic_ext, computes the
* requested stats, and serializes them back into the catalog.
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
*/
void
BuildRelationExtStatistics(Relation onerel, double totalrows,
int numrows, HeapTuple *rows,
int natts, VacAttrStats **vacattrstats)
{
Relation pg_stext;
ListCell *lc;
List *stats;
MemoryContext cxt;
MemoryContext oldcxt;
cxt = AllocSetContextCreate(CurrentMemoryContext, "stats ext",
ALLOCSET_DEFAULT_SIZES);
oldcxt = MemoryContextSwitchTo(cxt);
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
pg_stext = heap_open(StatisticExtRelationId, RowExclusiveLock);
stats = fetch_statentries_for_relation(pg_stext, RelationGetRelid(onerel));
foreach(lc, stats)
{
StatExtEntry *stat = (StatExtEntry *) lfirst(lc);
MVNDistinct *ndistinct = NULL;
MVDependencies *dependencies = NULL;
VacAttrStats **stats;
ListCell *lc2;
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
/*
* Check if we can build these stats based on the column analyzed. If
* not, report this fact (except in autovacuum) and move on.
*/
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
stats = lookup_var_attr_stats(onerel, stat->columns,
natts, vacattrstats);
if (!stats)
{
if (!IsAutoVacuumWorkerProcess())
ereport(WARNING,
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
errmsg("statistics object \"%s.%s\" could not be computed for relation \"%s.%s\"",
stat->schema, stat->name,
get_namespace_name(onerel->rd_rel->relnamespace),
RelationGetRelationName(onerel)),
errtable(onerel)));
continue;
}
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
/* check allowed number of dimensions */
Assert(bms_num_members(stat->columns) >= 2 &&
bms_num_members(stat->columns) <= STATS_MAX_DIMENSIONS);
/* compute statistic of each requested type */
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
foreach(lc2, stat->types)
{
char t = (char) lfirst_int(lc2);
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
if (t == STATS_EXT_NDISTINCT)
ndistinct = statext_ndistinct_build(totalrows, numrows, rows,
stat->columns, stats);
else if (t == STATS_EXT_DEPENDENCIES)
dependencies = statext_dependencies_build(numrows, rows,
stat->columns, stats);
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
}
/* store the statistics in the catalog */
statext_store(pg_stext, stat->statOid, ndistinct, dependencies, stats);
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
}
heap_close(pg_stext, RowExclusiveLock);
MemoryContextSwitchTo(oldcxt);
MemoryContextDelete(cxt);
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
}
/*
* statext_is_kind_built
* Is this stat kind built in the given pg_statistic_ext tuple?
*/
bool
statext_is_kind_built(HeapTuple htup, char type)
{
AttrNumber attnum;
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
switch (type)
{
case STATS_EXT_NDISTINCT:
attnum = Anum_pg_statistic_ext_stxndistinct;
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
break;
case STATS_EXT_DEPENDENCIES:
attnum = Anum_pg_statistic_ext_stxdependencies;
break;
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
default:
elog(ERROR, "unexpected statistics type requested: %d", type);
}
return !heap_attisnull(htup, attnum);
}
/*
* Return a list (of StatExtEntry) of statistics objects for the given relation.
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
*/
static List *
fetch_statentries_for_relation(Relation pg_statext, Oid relid)
{
SysScanDesc scan;
ScanKeyData skey;
HeapTuple htup;
List *result = NIL;
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
/*
* Prepare to scan pg_statistic_ext for entries having stxrelid = this
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
* rel.
*/
ScanKeyInit(&skey,
Anum_pg_statistic_ext_stxrelid,
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(relid));
scan = systable_beginscan(pg_statext, StatisticExtRelidIndexId, true,
NULL, 1, &skey);
while (HeapTupleIsValid(htup = systable_getnext(scan)))
{
StatExtEntry *entry;
Datum datum;
bool isnull;
int i;
ArrayType *arr;
char *enabled;
Form_pg_statistic_ext staForm;
entry = palloc0(sizeof(StatExtEntry));
entry->statOid = HeapTupleGetOid(htup);
staForm = (Form_pg_statistic_ext) GETSTRUCT(htup);
entry->schema = get_namespace_name(staForm->stxnamespace);
entry->name = pstrdup(NameStr(staForm->stxname));
for (i = 0; i < staForm->stxkeys.dim1; i++)
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
{
entry->columns = bms_add_member(entry->columns,
staForm->stxkeys.values[i]);
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
}
/* decode the stxkind char array into a list of chars */
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
datum = SysCacheGetAttr(STATEXTOID, htup,
Anum_pg_statistic_ext_stxkind, &isnull);
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
Assert(!isnull);
arr = DatumGetArrayTypeP(datum);
if (ARR_NDIM(arr) != 1 ||
ARR_HASNULL(arr) ||
ARR_ELEMTYPE(arr) != CHAROID)
elog(ERROR, "stxkind is not a 1-D char array");
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
enabled = (char *) ARR_DATA_PTR(arr);
for (i = 0; i < ARR_DIMS(arr)[0]; i++)
{
Assert((enabled[i] == STATS_EXT_NDISTINCT) ||
(enabled[i] == STATS_EXT_DEPENDENCIES));
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
entry->types = lappend_int(entry->types, (int) enabled[i]);
}
result = lappend(result, entry);
}
systable_endscan(scan);
return result;
}
/*
* Using 'vacatts' of size 'nvacatts' as input data, return a newly built
* VacAttrStats array which includes only the items corresponding to
* attributes indicated by 'stxkeys'. If we don't have all of the per column
* stats available to compute the extended stats, then we return NULL to indicate
* to the caller that the stats should not be built.
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
*/
static VacAttrStats **
lookup_var_attr_stats(Relation rel, Bitmapset *attrs,
int nvacatts, VacAttrStats **vacatts)
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
{
int i = 0;
int x = -1;
VacAttrStats **stats;
stats = (VacAttrStats **)
palloc(bms_num_members(attrs) * sizeof(VacAttrStats *));
/* lookup VacAttrStats info for the requested columns (same attnum) */
while ((x = bms_next_member(attrs, x)) >= 0)
{
int j;
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
stats[i] = NULL;
for (j = 0; j < nvacatts; j++)
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
{
if (x == vacatts[j]->tupattnum)
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
{
stats[i] = vacatts[j];
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
break;
}
}
if (!stats[i])
{
/*
* Looks like stats were not gathered for one of the columns
* required. We'll be unable to build the extended stats without
* this column.
*/
pfree(stats);
return NULL;
}
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
/*
* Sanity check that the column is not dropped - stats should have
* been removed in this case.
*/
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
Assert(!stats[i]->attr->attisdropped);
i++;
}
return stats;
}
/*
* statext_store
* Serializes the statistics and stores them into the pg_statistic_ext tuple.
*/
static void
statext_store(Relation pg_stext, Oid statOid,
MVNDistinct *ndistinct, MVDependencies *dependencies,
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
VacAttrStats **stats)
{
HeapTuple stup,
oldtup;
Datum values[Natts_pg_statistic_ext];
bool nulls[Natts_pg_statistic_ext];
bool replaces[Natts_pg_statistic_ext];
memset(nulls, true, sizeof(nulls));
memset(replaces, false, sizeof(replaces));
memset(values, 0, sizeof(values));
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
/*
* Construct a new pg_statistic_ext tuple, replacing the calculated stats.
*/
if (ndistinct != NULL)
{
bytea *data = statext_ndistinct_serialize(ndistinct);
nulls[Anum_pg_statistic_ext_stxndistinct - 1] = (data == NULL);
values[Anum_pg_statistic_ext_stxndistinct - 1] = PointerGetDatum(data);
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
}
if (dependencies != NULL)
{
bytea *data = statext_dependencies_serialize(dependencies);
nulls[Anum_pg_statistic_ext_stxdependencies - 1] = (data == NULL);
values[Anum_pg_statistic_ext_stxdependencies - 1] = PointerGetDatum(data);
}
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
/* always replace the value (either by bytea or NULL) */
replaces[Anum_pg_statistic_ext_stxndistinct - 1] = true;
replaces[Anum_pg_statistic_ext_stxdependencies - 1] = true;
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
/* there should already be a pg_statistic_ext tuple */
oldtup = SearchSysCache1(STATEXTOID, ObjectIdGetDatum(statOid));
if (!HeapTupleIsValid(oldtup))
elog(ERROR, "cache lookup failed for statistics object %u", statOid);
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
/* replace it */
stup = heap_modify_tuple(oldtup,
RelationGetDescr(pg_stext),
values,
nulls,
replaces);
ReleaseSysCache(oldtup);
CatalogTupleUpdate(pg_stext, &stup->t_self, stup);
heap_freetuple(stup);
}
/* initialize multi-dimensional sort */
MultiSortSupport
multi_sort_init(int ndims)
{
MultiSortSupport mss;
Assert(ndims >= 2);
mss = (MultiSortSupport) palloc0(offsetof(MultiSortSupportData, ssup)
2017-06-21 20:39:04 +02:00
+ sizeof(SortSupportData) * ndims);
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
mss->ndims = ndims;
return mss;
}
/*
* Prepare sort support info using the given sort operator
* at the position 'sortdim'
*/
void
multi_sort_add_dimension(MultiSortSupport mss, int sortdim, Oid oper)
{
SortSupport ssup = &mss->ssup[sortdim];
Implement multivariate n-distinct coefficients Add support for explicitly declared statistic objects (CREATE STATISTICS), allowing collection of statistics on more complex combinations that individual table columns. Companion commands DROP STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are added too. All this DDL has been designed so that more statistic types can be added later on, such as multivariate most-common-values and multivariate histograms between columns of a single table, leaving room for permitting columns on multiple tables, too, as well as expressions. This commit only adds support for collection of n-distinct coefficient on user-specified sets of columns in a single table. This is useful to estimate number of distinct groups in GROUP BY and DISTINCT clauses; estimation errors there can cause over-allocation of memory in hashed aggregates, for instance, so it's a worthwhile problem to solve. A new special pseudo-type pg_ndistinct is used. (num-distinct estimation was deemed sufficiently useful by itself that this is worthwhile even if no further statistic types are added immediately; so much so that another version of essentially the same functionality was submitted by Kyotaro Horiguchi: https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp though this commit does not use that code.) Author: Tomas Vondra. Some code rework by Álvaro. Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes, Ideriha Takeshi Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.cz https://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
2017-03-24 18:06:10 +01:00
ssup->ssup_cxt = CurrentMemoryContext;
ssup->ssup_collation = DEFAULT_COLLATION_OID;
ssup->ssup_nulls_first = false;
ssup->ssup_cxt = CurrentMemoryContext;
PrepareSortSupportFromOrderingOp(oper, ssup);
}
/* compare all the dimensions in the selected order */
int
multi_sort_compare(const void *a, const void *b, void *arg)
{
MultiSortSupport mss = (MultiSortSupport) arg;
SortItem *ia = (SortItem *) a;
SortItem *ib = (SortItem *) b;
int i;
for (i = 0; i < mss->ndims; i++)
{
int compare;
compare = ApplySortComparator(ia->values[i], ia->isnull[i],
ib->values[i], ib->isnull[i],
&mss->ssup[i]);
if (compare != 0)
return compare;
}
/* equal by default */
return 0;
}
/* compare selected dimension */
int
multi_sort_compare_dim(int dim, const SortItem *a, const SortItem *b,
MultiSortSupport mss)
{
return ApplySortComparator(a->values[dim], a->isnull[dim],
b->values[dim], b->isnull[dim],
&mss->ssup[dim]);
}
int
multi_sort_compare_dims(int start, int end,
const SortItem *a, const SortItem *b,
MultiSortSupport mss)
{
int dim;
for (dim = start; dim <= end; dim++)
{
int r = ApplySortComparator(a->values[dim], a->isnull[dim],
b->values[dim], b->isnull[dim],
&mss->ssup[dim]);
if (r != 0)
return r;
}
return 0;
}
/*
* has_stats_of_kind
* Check whether the list contains statistic of a given kind
*/
bool
has_stats_of_kind(List *stats, char requiredkind)
{
ListCell *l;
foreach(l, stats)
{
StatisticExtInfo *stat = (StatisticExtInfo *) lfirst(l);
if (stat->kind == requiredkind)
return true;
}
return false;
}
/*
* choose_best_statistics
* Look for and return statistics with the specified 'requiredkind' which
* have keys that match at least two of the given attnums. Return NULL if
* there's no match.
*
* The current selection criteria is very simple - we choose the statistics
* object referencing the most of the requested attributes, breaking ties
* in favor of objects with fewer keys overall.
*
* XXX if multiple statistics objects tie on both criteria, then which object
* is chosen depends on the order that they appear in the stats list. Perhaps
* further tiebreakers are needed.
*/
StatisticExtInfo *
choose_best_statistics(List *stats, Bitmapset *attnums, char requiredkind)
{
ListCell *lc;
StatisticExtInfo *best_match = NULL;
int best_num_matched = 2; /* goal #1: maximize */
int best_match_keys = (STATS_MAX_DIMENSIONS + 1); /* goal #2: minimize */
foreach(lc, stats)
{
StatisticExtInfo *info = (StatisticExtInfo *) lfirst(lc);
int num_matched;
int numkeys;
Bitmapset *matched;
/* skip statistics that are not of the correct type */
if (info->kind != requiredkind)
continue;
/* determine how many attributes of these stats can be matched to */
matched = bms_intersect(attnums, info->keys);
num_matched = bms_num_members(matched);
bms_free(matched);
/*
* save the actual number of keys in the stats so that we can choose
* the narrowest stats with the most matching keys.
*/
numkeys = bms_num_members(info->keys);
/*
* Use this object when it increases the number of matched clauses or
* when it matches the same number of attributes but these stats have
* fewer keys than any previous match.
*/
if (num_matched > best_num_matched ||
(num_matched == best_num_matched && numkeys < best_match_keys))
{
best_match = info;
best_num_matched = num_matched;
best_match_keys = numkeys;
}
}
return best_match;
}