pgsql/ oc/src/sgml/catalogs.sgml oc/src/sgml/i ...

CVSROOT: /home/projects/pgsql/cvsroot
Module name: pgsql
Changes by: tgl(at)hub(dot)org 01/05/06 20:43:27

Modified files:
doc/src/sgml : catalogs.sgml indices.sgml reference.sgml
xoper.sgml
doc/src/sgml/ref: allfiles.sgml alter_table.sgml vacuum.sgml
src/backend/access/common: tupdesc.c
src/backend/access/gist: gist.c
src/backend/access/hash: hash.c
src/backend/access/heap: tuptoaster.c
src/backend/access/nbtree: nbtree.c
src/backend/access/rtree: rtree.c
src/backend/catalog: genbki.sh heap.c index.c
src/backend/commands: analyze.c command.c vacuum.c
src/backend/executor: nodeSort.c
src/backend/nodes: copyfuncs.c equalfuncs.c readfuncs.c
src/backend/optimizer/path: costsize.c joinpath.c
src/backend/optimizer/plan: createplan.c initsplan.c planner.c
src/backend/optimizer/prep: prepunion.c
src/backend/optimizer/util: pathnode.c plancat.c
src/backend/parser: analyze.c gram.y keywords.c parse_relation.c
src/backend/tcop: utility.c
src/backend/utils/adt: selfuncs.c
src/backend/utils/cache: lsyscache.c syscache.c
src/backend/utils/sort: tuplesort.c
src/include/access: tuptoaster.h
src/include/catalog: catversion.h heap.h index.h indexing.h
pg_attribute.h pg_class.h pg_statistic.h
src/include/commands: command.h vacuum.h
src/include : config.h.in
src/include/nodes: execnodes.h parsenodes.h primnodes.h
relation.h
src/include/optimizer: cost.h pathnode.h
src/include/utils: lsyscache.h syscache.h tuplesort.h
src/interfaces/ecpg/preproc: keywords.c preproc.y
src/test/regress/expected: oidjoins.out opr_sanity.out
src/test/regress/sql: oidjoins.sql
Added files:
doc/src/sgml/ref: analyze.sgml

Log message:
Rewrite of planner statistics-gathering code. ANALYZE is now available as
a separate statement (though it can still be invoked as part of VACUUM, too).
pg_statistic redesigned to be more flexible about what statistics are
stored. ANALYZE now collects a list of several of the most common values,
not just one, plus a histogram (not just the min and max values). Random
sampling is used to make the process reasonably fast even on very large
tables. The number of values and histogram bins collected is now
user-settable via an ALTER TABLE command.

There is more still to do; the new stats are not being used everywhere
they could be in the planner. But the remaining changes for this project
should be localized, and the behavior is already better than before.

A not-very-related change is that sorting now makes use of btree comparison
routines if it can find one, rather than invoking '<' twice.