]> granicus.if.org Git - postgresql/log
postgresql
25 years agoFurther planner/optimizer cleanups. Move all set_tlist_references
Tom Lane [Sun, 22 Aug 1999 20:15:04 +0000 (20:15 +0000)]
Further planner/optimizer cleanups.  Move all set_tlist_references
and fix_opids processing to a single recursive pass over the plan tree
executed at the very tail end of planning, rather than haphazardly here
and there at different places.  Now that tlist Vars do not get modified
until the very end, it's possible to get rid of the klugy var_equal and
match_varid partial-matching routines, and just use plain equal()
throughout the optimizer.  This is a step towards allowing merge and
hash joins to be done on expressions instead of only Vars ...

25 years agoMajor revision of sort-node handling: push knowledge of query
Tom Lane [Sat, 21 Aug 1999 03:49:17 +0000 (03:49 +0000)]
Major revision of sort-node handling: push knowledge of query
sort order down into planner, instead of handling it only at the very top
level of the planner.  This fixes many things.  An explicit sort is now
avoided if there is a cheaper alternative (typically an indexscan) not
only for ORDER BY, but also for the internal sort of GROUP BY.  It works
even when there is no other reason (such as a WHERE condition) to consider
the indexscan.  It works for indexes on functions.  It works for indexes
on functions, backwards.  It's just so cool...

CAUTION: I have changed the representation of SortClause nodes, therefore
THIS UPDATE BREAKS STORED RULES.  You will need to initdb.

25 years agoCleanups for int8: guard against null inputs in comparison
Tom Lane [Sat, 21 Aug 1999 03:06:58 +0000 (03:06 +0000)]
Cleanups for int8: guard against null inputs in comparison
operators (and some other places), fix rangechecks in int8 to int4
conversion (same problem we recently figured out in pg_atoi).

25 years agoOoops ... I had left some test coding in selfuncs.c that
Tom Lane [Sat, 21 Aug 1999 00:56:18 +0000 (00:56 +0000)]
Ooops ... I had left some test coding in selfuncs.c that
failed on 'field < textconstant' ...

25 years agoUpdate TODO list.
Bruce Momjian [Wed, 18 Aug 1999 15:11:52 +0000 (15:11 +0000)]
Update TODO list.

25 years agoOld multi-byte bug. Forgot to rename #ifdef MB to #ifdef MULTIBYTE
Tatsuo Ishii [Wed, 18 Aug 1999 13:04:45 +0000 (13:04 +0000)]
Old multi-byte bug. Forgot to rename #ifdef MB to #ifdef MULTIBYTE
Now SET NAMES working again...

25 years agoRemove extraneous SeqScan node that make_noname was inserting
Tom Lane [Wed, 18 Aug 1999 04:15:16 +0000 (04:15 +0000)]
Remove extraneous SeqScan node that make_noname was inserting
above a Sort or Materialize node.  As far as I can tell, the only place
that actually needed that was set_tlist_references, which was being lazy
about checking to see if it had a noname node to fix or not...

25 years agoAdd script that runs the regression tests with all valid
Tom Lane [Tue, 17 Aug 1999 21:21:22 +0000 (21:21 +0000)]
Add script that runs the regression tests with all valid
combinations of query-plan-type backend options.  Good for testing
planner/optimizer.  Tedious, though.

25 years agoAdd missing docs semicolon.
Bruce Momjian [Tue, 17 Aug 1999 17:31:11 +0000 (17:31 +0000)]
Add missing docs semicolon.

25 years agoEXPLAIN didn't know about 'Materialize' plan nodes.
Tom Lane [Mon, 16 Aug 1999 23:47:23 +0000 (23:47 +0000)]
EXPLAIN didn't know about 'Materialize' plan nodes.

25 years agoAssign sort keys properly when there are duplicate entries in
Tom Lane [Mon, 16 Aug 1999 23:07:20 +0000 (23:07 +0000)]
Assign sort keys properly when there are duplicate entries in
pathkey list --- corrects misbehavior seen with multiple mergejoin clauses
mentioning same variable.

25 years agoTypo fix
Bruce Momjian [Mon, 16 Aug 1999 20:32:34 +0000 (20:32 +0000)]
Typo fix

25 years ago I've sent 3 mails to pgsql-patches. There are two files, one for doc
Bruce Momjian [Mon, 16 Aug 1999 20:27:19 +0000 (20:27 +0000)]
   I've sent 3 mails to pgsql-patches. There are two files, one for doc
and
for src/data directories, and one minor patch for doc/README.locale.
   Please apply.

Oleg.

25 years agoFix for perl5 on BSD/OS.
Bruce Momjian [Mon, 16 Aug 1999 20:12:12 +0000 (20:12 +0000)]
Fix for perl5 on BSD/OS.

25 years agoChange >& to > and 2>&1.
Bruce Momjian [Mon, 16 Aug 1999 19:59:25 +0000 (19:59 +0000)]
Change >& to > and 2>&1.

25 years agoAllow BSD yacc and bison to compile pl code.
Bruce Momjian [Mon, 16 Aug 1999 19:57:21 +0000 (19:57 +0000)]
Allow BSD yacc and bison to compile pl code.

25 years agoMajor planner/optimizer revision: get rid of PathOrder node type,
Tom Lane [Mon, 16 Aug 1999 02:17:58 +0000 (02:17 +0000)]
Major planner/optimizer revision: get rid of PathOrder node type,
store all ordering information in pathkeys lists (which are now lists of
lists of PathKeyItem nodes, not just lists of lists of vars).  This was
a big win --- the code is smaller and IMHO more understandable than it
was, even though it handles more cases.  I believe the node changes will
not force an initdb for anyone; planner nodes don't show up in stored
rules.

25 years agoSmall updates to #include lists for pending optimizer checkin.
Tom Lane [Mon, 16 Aug 1999 02:10:13 +0000 (02:10 +0000)]
Small updates to #include lists for pending optimizer checkin.

25 years agoMove funcid_get_rettype() to lsyscache.
Tom Lane [Mon, 16 Aug 1999 02:08:59 +0000 (02:08 +0000)]
Move funcid_get_rettype() to lsyscache.

25 years agoAdd get_func_rettype() to general-use lsyscache routines,
Tom Lane [Mon, 16 Aug 1999 02:06:25 +0000 (02:06 +0000)]
Add get_func_rettype() to general-use lsyscache routines,
since it's now needed in both optimizer and parser.

25 years agoFix for Win32 making problem with MB enabled.
Tatsuo Ishii [Mon, 16 Aug 1999 02:04:05 +0000 (02:04 +0000)]
Fix for Win32 making problem with MB enabled.
Patches created by Hiroki Kataoka.

25 years agoRepair the check for redundant UNIQUE and PRIMARY KEY indices.
Thomas G. Lockhart [Sun, 15 Aug 1999 06:46:49 +0000 (06:46 +0000)]
Repair the check for redundant UNIQUE and PRIMARY KEY indices.
Also, improve it so that it checks for multi-column constraints.
Thanks to Mark Dalphin <mdalphin@amgen.com> for reporting the problem.

25 years agoLispUnion fix changes output ordering for inheritance tests.
Tom Lane [Sat, 14 Aug 1999 19:30:21 +0000 (19:30 +0000)]
LispUnion fix changes output ordering for inheritance tests.

25 years agoLispUnion routine didn't generate a proper union: anytime
Tom Lane [Sat, 14 Aug 1999 19:29:35 +0000 (19:29 +0000)]
LispUnion routine didn't generate a proper union: anytime
l2 contained more than one entry, there would be duplicates in the output
list.  Miscellaneous code beautification in other routines, too.

25 years agoUpdate TODO list.
Bruce Momjian [Fri, 13 Aug 1999 16:56:04 +0000 (16:56 +0000)]
Update TODO list.

25 years agoUpdate comments about pathkeys.
Tom Lane [Fri, 13 Aug 1999 01:17:16 +0000 (01:17 +0000)]
Update comments about pathkeys.

25 years agoClean up optimizer's handling of indexscan quals that need to be
Tom Lane [Thu, 12 Aug 1999 04:32:54 +0000 (04:32 +0000)]
Clean up optimizer's handling of indexscan quals that need to be
commuted (ie, the index var appears on the right).  These are now handled
the same way as merge and hash join quals that need to be commuted: the
actual reversing of the clause only happens if we actually choose the path
and generate a plan from it.  Furthermore, the clause is only reversed in
the 'indexqual' field of the plan, not in the 'indxqualorig' field.  This
allows the clause to still be recognized and removed from qpquals of upper
level join plans.  Also, simplify and generalize match_clause_to_indexkey;
now it recognizes binary-compatible indexes for join as well as restriction
clauses.

25 years agoAdd commentary to show that even though ExecInitIndexScan()
Tom Lane [Thu, 12 Aug 1999 00:42:43 +0000 (00:42 +0000)]
Add commentary to show that even though ExecInitIndexScan()
contains much code that looks like it will handle indexquals with the index
key on either side of the operator, in fact indexquals must have the index
key on the left because of limitations of the ScanKey machinery.  Perhaps
someone will be motivated to fix that someday...

25 years agoUpdate TODO list.
Bruce Momjian [Tue, 10 Aug 1999 16:56:54 +0000 (16:56 +0000)]
Update TODO list.

25 years agoMinor cleanups and code beautification; eliminate some
Tom Lane [Tue, 10 Aug 1999 03:00:15 +0000 (03:00 +0000)]
Minor cleanups and code beautification; eliminate some
routines that are now dead code.

25 years agoRevise create_nestloop_node's handling of inner indexscan to
Tom Lane [Tue, 10 Aug 1999 02:58:56 +0000 (02:58 +0000)]
Revise create_nestloop_node's handling of inner indexscan to
work under a wider range of scenarios than it did --- it formerly did not
handle a multi-pass inner scan, nor cases in which the inner scan's
indxqualorig or non-index qual contained outer var references.  I am not
sure that these limitations could be hit in the existing optimizer, but
they need to be fixed for future expansion.

25 years ago> > Prevent sorting if result is already sorted
Bruce Momjian [Mon, 9 Aug 1999 06:20:27 +0000 (06:20 +0000)]
> >        Prevent sorting if result is already sorted
> >
> > was implemented by Jan Wieck.
> > His work is for ascending order cases.
> >
> > Here is a patch to prevent sorting also in descending
> > order cases.
> > Because I had already changed _bt_first() to position
> > backward correctly before v6.5,this patch would work.
> >
Hiroshi Inoue
Inoue@tpf.co.jp

25 years agoClean up tlist.c tree-walking routines with
Tom Lane [Mon, 9 Aug 1999 05:34:13 +0000 (05:34 +0000)]
Clean up tlist.c tree-walking routines with
expression_tree_mutator.

25 years agoStore -1 in attdisbursion to signal 'no duplicates in column'.
Tom Lane [Mon, 9 Aug 1999 03:16:47 +0000 (03:16 +0000)]
Store -1 in attdisbursion to signal 'no duplicates in column'.
Centralize att_disbursion readout logic.

25 years agoMove get_attdisbursion to lsyscache. Clean up get_typdefault.
Tom Lane [Mon, 9 Aug 1999 03:13:31 +0000 (03:13 +0000)]
Move get_attdisbursion to lsyscache.  Clean up get_typdefault.

25 years agoUpdate comments about attdisbursion. NO code change.
Tom Lane [Mon, 9 Aug 1999 02:45:56 +0000 (02:45 +0000)]
Update comments about attdisbursion.  NO code change.

25 years agoRe-use free space on index pages with duplicates.
Vadim B. Mikheev [Mon, 9 Aug 1999 01:39:19 +0000 (01:39 +0000)]
Re-use free space on index pages with duplicates.

25 years agoRewrite fix_indxqual_references, which was entirely bogus for
Tom Lane [Mon, 9 Aug 1999 01:01:42 +0000 (01:01 +0000)]
Rewrite fix_indxqual_references, which was entirely bogus for
multi-scan indexscan plans; it tried to use the same table-to-index
attribute mapping for all the scans, even if they used different indexes.
It would klugily work as long as OR indexquals never used multikey indexes,
but that's not likely to hold up much longer...

25 years agoClean up routines in setrefs.c by replacing individual tree
Tom Lane [Mon, 9 Aug 1999 00:56:05 +0000 (00:56 +0000)]
Clean up routines in setrefs.c by replacing individual tree
walking logic with expression_tree_walker/mutator calls.

25 years agoCreate a standardized expression_tree_mutator support routine
Tom Lane [Mon, 9 Aug 1999 00:51:26 +0000 (00:51 +0000)]
Create a standardized expression_tree_mutator support routine
to go along with expression_tree_walker.  (_walker is not suitable for
routines that need to alter the tree structure significantly.)  Other minor
cleanups in clauses.c.

25 years agoreplace #Include .
Bruce Momjian [Mon, 9 Aug 1999 00:08:52 +0000 (00:08 +0000)]
replace #Include .

25 years agoFix nbtree's failure to clear BTScans list during xact abort.
Tom Lane [Sun, 8 Aug 1999 20:12:52 +0000 (20:12 +0000)]
Fix nbtree's failure to clear BTScans list during xact abort.
Also, move responsibility for calling vc_abort into main xact.c list of
things-to-call-at-abort.  What in the world was it doing down inside of
TransactionIdAbort()?

25 years agoFor a unique-key attribute (no duplicate values), vacuum analyze
Tom Lane [Sun, 8 Aug 1999 17:13:10 +0000 (17:13 +0000)]
For a unique-key attribute (no duplicate values), vacuum analyze
was recording a disbursion of 0, not the correct value 1/numberOfRows.

25 years agoRemove old man pages. All info is in new man pages or other docs.
Thomas G. Lockhart [Sun, 8 Aug 1999 15:21:50 +0000 (15:21 +0000)]
Remove old man pages. All info is in new man pages or other docs.

25 years agoUse sgml/Makefile to generate man sections, so just "make man" there.
Thomas G. Lockhart [Sun, 8 Aug 1999 15:10:21 +0000 (15:10 +0000)]
Use sgml/Makefile to generate man sections, so just "make man" there.

25 years agoSwitch from old man pages to new ones generated from sgml.
Thomas G. Lockhart [Sun, 8 Aug 1999 15:09:04 +0000 (15:09 +0000)]
Switch from old man pages to new ones generated from sgml.

25 years agoRemove emacs properties block from end of file.
Thomas G. Lockhart [Sun, 8 Aug 1999 15:05:22 +0000 (15:05 +0000)]
Remove emacs properties block from end of file.

25 years agoUpdate references to char2 type by using char(2).
Thomas G. Lockhart [Sun, 8 Aug 1999 04:21:33 +0000 (04:21 +0000)]
Update references to char2 type by using char(2).
Thanks to Garr Updegraff <garru@uci.edu> for the tip.

25 years agoAllow nsgmls to validate input by defining the document type.
Thomas G. Lockhart [Sun, 8 Aug 1999 04:18:29 +0000 (04:18 +0000)]
Allow nsgmls to validate input by defining the document type.
Use refentry.sgml as a generic header for man pages through nsgmls.

25 years agoFix so BSD yacc works on file.
Bruce Momjian [Sat, 7 Aug 1999 04:25:02 +0000 (04:25 +0000)]
Fix so BSD yacc works on file.

25 years agoFix markup to allow clean man pages.
Thomas G. Lockhart [Fri, 6 Aug 1999 13:50:31 +0000 (13:50 +0000)]
Fix markup to allow clean man pages.
Add new pages for programming languages and tcl shells.

25 years agoFix cross-reference markup so that only the *title* of the Operators
Thomas G. Lockhart [Fri, 6 Aug 1999 13:48:38 +0000 (13:48 +0000)]
Fix cross-reference markup so that only the *title* of the Operators
 chapter is included, not the chapter itself.
Thanks to Evelyn Mitchell <efm@tummy.com> for pointing it out.

25 years agoRemove explicit references to ref/ path in file names; use vpath instead.
Thomas G. Lockhart [Fri, 6 Aug 1999 13:46:17 +0000 (13:46 +0000)]
Remove explicit references to ref/ path in file names; use vpath instead.
Fix rules for man pages to ensure double-pass to get cross references.
Add a few new man pages.

25 years agoTry to clarify characteristics of the SERIAL type.
Thomas G. Lockhart [Fri, 6 Aug 1999 13:43:42 +0000 (13:43 +0000)]
Try to clarify characteristics of the SERIAL type.
Fix source indenting, which does not affect output.
Note: still need docs on NUMERIC and DECIMAL
 (and let's not talk about regression tests :()

25 years agoRevise generation of hashjoin paths: generate one path per
Tom Lane [Fri, 6 Aug 1999 04:00:17 +0000 (04:00 +0000)]
Revise generation of hashjoin paths: generate one path per
hashjoinable clause, not one path for a randomly-chosen element of each
set of clauses with the same join operator.  That is, if you wrote
   SELECT ... WHERE t1.f1 = t2.f2 and t1.f3 = t2.f4,
and both '=' ops were the same opcode (say, all four fields are int4),
then the system would either consider hashing on f1=f2 or on f3=f4,
but it would *not* consider both possibilities.  Boo hiss.
Also, revise estimation of hashjoin costs to include a penalty when the
inner join var has a high disbursion --- ie, the most common value is
pretty common.  This tends to lead to badly skewed hash bucket occupancy
and way more comparisons than you'd expect on average.
I imagine that the cost calculation still needs tweaking, but at least
it generates a more reasonable plan than before on George Young's example.

25 years agoparse_coerce fix allows early simplification of one
Tom Lane [Thu, 5 Aug 1999 02:35:29 +0000 (02:35 +0000)]
parse_coerce fix allows early simplification of one
constant-coercion expression in the rules test.

25 years agoRevise parse_coerce() to handle coercion of int and float
Tom Lane [Thu, 5 Aug 1999 02:33:54 +0000 (02:33 +0000)]
Revise parse_coerce() to handle coercion of int and float
constants, not only string constants, at parse time.  Get rid of
parser_typecast2(), which is bogus and redundant...

25 years agoUpdate TODO list.
Bruce Momjian [Thu, 5 Aug 1999 01:53:35 +0000 (01:53 +0000)]
Update TODO list.

25 years agoFix ELF test so it doesn't spit up on all non-ELF systems...
Tom Lane [Tue, 3 Aug 1999 00:09:32 +0000 (00:09 +0000)]
Fix ELF test so it doesn't spit up on all non-ELF systems...
use Autoconf-approved method of testing for predefined symbols, and move
it down to where we know what compiler to run and how to run it.

25 years agoAdd a VACUUM command in hopes of making pg_upgrade usable again
Tom Lane [Mon, 2 Aug 1999 22:34:53 +0000 (22:34 +0000)]
Add a VACUUM command in hopes of making pg_upgrade usable again
in MVCC environment.  I do not trust this until Vadim says it's OK...

25 years ago*** empty log message ***
Michael Meskes [Mon, 2 Aug 1999 04:59:18 +0000 (04:59 +0000)]
*** empty log message ***

25 years agoFurther selectivity-estimation work. Speed up eqsel()
Tom Lane [Mon, 2 Aug 1999 02:05:41 +0000 (02:05 +0000)]
Further selectivity-estimation work.  Speed up eqsel()
(it should just call the given operator, not look up an = operator).
Fix intltsel() so that all numeric data types are converted to double
before trying to estimate where the given comparison value is in the
known range of column values.  intltsel() still needs work, or replacement,
for non-numeric data types ... but for nonintegral numeric types it
should now be delivering reasonable estimates.

25 years agoback out last...
Marc G. Fournier [Mon, 2 Aug 1999 01:42:57 +0000 (01:42 +0000)]
back out last...

25 years agotesting somethign...
Marc G. Fournier [Mon, 2 Aug 1999 01:40:30 +0000 (01:40 +0000)]
testing somethign...

25 years agoautoconf
Bruce Momjian [Sun, 1 Aug 1999 16:36:57 +0000 (16:36 +0000)]
autoconf

25 years agoI didn't see any further discussion so here is, I hope, a clean fix to
Bruce Momjian [Sun, 1 Aug 1999 16:30:05 +0000 (16:30 +0000)]
I didn't see any further discussion so here is, I hope, a clean fix to
configure.in to determine if a system is ELF or not.  Note that some
of the tests earlier may be redundant but I took the safest route.

D'Arcy J.M. Cain

25 years agoUpdate NT readme.
Bruce Momjian [Sun, 1 Aug 1999 14:48:50 +0000 (14:48 +0000)]
Update NT readme.

25 years agoFirst step in fixing selectivity-estimation code. eqsel and
Tom Lane [Sun, 1 Aug 1999 04:54:25 +0000 (04:54 +0000)]
First step in fixing selectivity-estimation code. eqsel and
neqsel now behave as per my suggestions in pghackers a few days ago.
selectivity for < > <= >= should work OK for integral types as well, but
still need work for nonintegral types.  Since these routines have never
actually executed before :-(, this may result in some significant changes
in the optimizer's choices of execution plans.  Let me know if you see
any serious misbehavior.
CAUTION: THESE CHANGES REQUIRE INITDB.  pg_statistic table has changed.

25 years agoAdd another ORDER BY to rules test; got tired of it 'failing'
Tom Lane [Sun, 1 Aug 1999 04:41:32 +0000 (04:41 +0000)]
Add another ORDER BY to rules test; got tired of it 'failing'
every time I tweak the optimizer...

25 years agoRe-enable pg_upgrade, after adding checks that the source
Tom Lane [Sat, 31 Jul 1999 22:06:44 +0000 (22:06 +0000)]
Re-enable pg_upgrade, after adding checks that the source
and target databases are of versions it knows about.

25 years agoAdd some more verbiage to man page's directions for pg_upgrade.
Tom Lane [Sat, 31 Jul 1999 22:05:48 +0000 (22:05 +0000)]
Add some more verbiage to man page's directions for pg_upgrade.

25 years agoAdd comments for attdisbursion field --- NO code change.
Tom Lane [Sat, 31 Jul 1999 19:07:25 +0000 (19:07 +0000)]
Add comments for attdisbursion field --- NO code change.

25 years agoUpdate comments about clause selectivity estimation.
Tom Lane [Fri, 30 Jul 1999 22:34:19 +0000 (22:34 +0000)]
Update comments about clause selectivity estimation.

25 years agoMake usecatupd disabled for normal users, and allow normal users to
Bruce Momjian [Fri, 30 Jul 1999 18:09:49 +0000 (18:09 +0000)]
Make usecatupd disabled for normal users, and allow normal users to
update temp tables with this setting.

25 years agoFurther cleanups of indexqual processing: simplify control
Tom Lane [Fri, 30 Jul 1999 04:07:25 +0000 (04:07 +0000)]
Further cleanups of indexqual processing: simplify control
logic in indxpath.c, avoid generation of redundant indexscan paths for the
same relation and index.

25 years agoThere is one section that changed, concernign startup...the rest is just
Marc G. Fournier [Fri, 30 Jul 1999 04:04:53 +0000 (04:04 +0000)]
There is one section that changed, concernign startup...the rest is just
changes for v6.5->v6.5.1, so relatively harmless

25 years agov6.6's HISTORY file should reflect changes that went into all previous
Marc G. Fournier [Fri, 30 Jul 1999 03:58:20 +0000 (03:58 +0000)]
v6.6's HISTORY file should reflect changes that went into all previous
releases, including v6.5.1 ...

25 years agojust testing a script...
Marc G. Fournier [Fri, 30 Jul 1999 03:45:57 +0000 (03:45 +0000)]
just testing a script...

25 years agoRemove extra #endif
Bruce Momjian [Fri, 30 Jul 1999 03:45:44 +0000 (03:45 +0000)]
Remove extra #endif

25 years agoNothing changed, just testing cvslog ...
Marc G. Fournier [Fri, 30 Jul 1999 03:42:59 +0000 (03:42 +0000)]
Nothing changed, just testing cvslog ...

25 years agoFix description of psql flags -t and -q.
Bruce Momjian [Fri, 30 Jul 1999 00:59:50 +0000 (00:59 +0000)]
Fix description of psql flags -t and -q.

25 years agoFix coredump seen when doing mergejoin between indexed tables,
Tom Lane [Fri, 30 Jul 1999 00:56:17 +0000 (00:56 +0000)]
Fix coredump seen when doing mergejoin between indexed tables,
for example in the regression test database, try
select * from tenk1 t1, tenk1 t2 where t1.unique1 = t2.unique2;
6.5 has this same bug ...

25 years agoUpdate comments for create_indexscan_node().
Tom Lane [Fri, 30 Jul 1999 00:44:23 +0000 (00:44 +0000)]
Update comments for create_indexscan_node().

25 years agoRemove comment after optimization flag.
Bruce Momjian [Thu, 29 Jul 1999 15:39:05 +0000 (15:39 +0000)]
Remove comment after optimization flag.

25 years agoAdd support for Case exprs to fix_indxqual_references,
Tom Lane [Thu, 29 Jul 1999 02:48:05 +0000 (02:48 +0000)]
Add support for Case exprs to fix_indxqual_references,
so that Case works in WHERE join clauses.  Temporary patch --- this routine
is one of many that ought to be changed to use centralized expression-tree-
walking logic.

25 years agoAdd equal() funcs for Case nodes ... amazing we had not
Tom Lane [Thu, 29 Jul 1999 02:45:36 +0000 (02:45 +0000)]
Add equal() funcs for Case nodes ... amazing we had not
detected this omission before.  Miscellaneous other cleanups.

25 years agoUpdate TODO list.
Bruce Momjian [Wed, 28 Jul 1999 17:52:31 +0000 (17:52 +0000)]
Update TODO list.

25 years agoUpdate TODO list.
Bruce Momjian [Wed, 28 Jul 1999 17:51:41 +0000 (17:51 +0000)]
Update TODO list.

25 years agoAllow a_expr not just AexprConst in the right-hand list of
Tom Lane [Wed, 28 Jul 1999 17:39:38 +0000 (17:39 +0000)]
Allow a_expr not just AexprConst in the right-hand list of
IN and NOT IN operators.  Rewrite grotty implementation of IN-list
parsing ... look Ma, no global variable ...

25 years agoUpdate TODO list.
Bruce Momjian [Wed, 28 Jul 1999 00:09:49 +0000 (00:09 +0000)]
Update TODO list.

25 years agoCorrect bug in best_innerjoin(): it should check all the
Tom Lane [Tue, 27 Jul 1999 06:23:12 +0000 (06:23 +0000)]
Correct bug in best_innerjoin(): it should check all the
rels that the inner path needs to join to, but it was only checking for
the first one.  Failure could only have been observed with an OR-clause
that mentions 3 or more tables, and then only if the bogus path was
actually selected as cheapest ...

25 years agoMy recent optimizer changes caused a different plan to
Tom Lane [Tue, 27 Jul 1999 03:53:54 +0000 (03:53 +0000)]
My recent optimizer changes caused a different plan to
be picked for one of the complex joins in rules test ... leading to
a different output ordering ...

25 years agoFirst cut at doing LIKE/regex indexing optimization in
Tom Lane [Tue, 27 Jul 1999 03:51:11 +0000 (03:51 +0000)]
First cut at doing LIKE/regex indexing optimization in
optimizer rather than parser.  This has many advantages, such as not
getting fooled by chance uses of operator names ~ and ~~ (the operators
are identified by OID now), and not creating useless comparison operations
in contexts where the comparisons will not actually be used as indexquals.
The new code also recognizes exact-match LIKE and regex patterns, and
produces an = indexqual instead of >= and <=.

This change does NOT fix the problem with non-ASCII locales: the code
still doesn't know how to generate an upper bound indexqual for non-ASCII
collation order.  But it's no worse than before, just the same deficiency
in a different place...

Also, dike out loc_restrictinfo fields in Plan nodes.  These were doing
nothing useful in the absence of 'expensive functions' optimization,
and they took a considerable amount of processing to fill in.

25 years agoUpdate TODO list.
Bruce Momjian [Mon, 26 Jul 1999 20:12:33 +0000 (20:12 +0000)]
Update TODO list.

25 years agoUpdate TODO list.
Bruce Momjian [Mon, 26 Jul 1999 15:28:32 +0000 (15:28 +0000)]
Update TODO list.

25 years agoFurther work on planning of indexscans. Cleaned up interfaces
Tom Lane [Sun, 25 Jul 1999 23:07:26 +0000 (23:07 +0000)]
Further work on planning of indexscans.  Cleaned up interfaces
to index_selectivity so that it can be handed an indexqual clause list
rather than a bunch of assorted derivative data.

25 years agoRemove 'restrictinfojoinid' field from RestrictInfo nodes.
Tom Lane [Sun, 25 Jul 1999 17:53:27 +0000 (17:53 +0000)]
Remove 'restrictinfojoinid' field from RestrictInfo nodes.
The only place it was being used was as temporary storage in indxpath.c,
and the logic was wrong: the same restrictinfo node could get chosen to
carry the info for two different joins.  Right fix is to return a second
list of unjoined-relids parallel to the list of clause groups.

25 years agoClean up messy clause-selectivity code in clausesel.c; repair bug
Tom Lane [Sat, 24 Jul 1999 23:21:14 +0000 (23:21 +0000)]
Clean up messy clause-selectivity code in clausesel.c; repair bug
identified by Hiroshi (incorrect cost attributed to OR clauses
after multiple passes through set_rest_selec()).  I think the code
was trying to allow selectivities of OR subclauses to be passed in
from outside, but noplace was actually passing any useful data, and
set_rest_selec() was passing wrong data.

Restructure representation of "indexqual" in IndexPath nodes so that
it is the same as for indxqual in completed IndexScan nodes: namely,
a toplevel list with an entry for each pass of the index scan, having
sublists that are implicitly-ANDed index qual conditions for that pass.
You don't want to know what the old representation was :-(

Improve documentation of OR-clause indexscan functions.

Remove useless 'notclause' field from RestrictInfo nodes.  (This might
force an initdb for anyone who has stored rules containing RestrictInfos,
but I do not think that RestrictInfo ever appears in completed plans.)

25 years agoMinor code beautification, extensive improvement of
Tom Lane [Fri, 23 Jul 1999 03:34:49 +0000 (03:34 +0000)]
Minor code beautification, extensive improvement of
comments.  This file was full of obsolete and just plain wrong
commentary...

25 years agoExit cleanups I made yesterday caused pq_close() to be
Tom Lane [Fri, 23 Jul 1999 03:00:10 +0000 (03:00 +0000)]
Exit cleanups I made yesterday caused pq_close() to be
invoked during exit from a standalone backend, leading to core dump.
This is the cause of the recently reported initdb-time crash :-(.
Sorry folks...