the old JOIN_IN code, but antijoins are new functionality.) Teach the planner
to convert appropriate EXISTS and NOT EXISTS subqueries into semi and anti
joins respectively. Also, LEFT JOINs with suitable upper-level IS NULL
filters are recognized as being anti joins. Unify the InClauseInfo and
OuterJoinInfo infrastructure into "SpecialJoinInfo". With that change,
it becomes possible to associate a SpecialJoinInfo with every join attempt,
which permits some cleanup of join selectivity estimation. That needs to be
taken much further than this patch does, but the next step is to change the
API for oprjoin selectivity functions, which seems like material for a
separate patch. So for the moment the output size estimates for semi and
especially anti joins are quite bogus.
hashtable entries for tuples that are found only in the second input: they
can never contribute to the output. Furthermore, this implies that the
planner should endeavor to put first the smaller (in number of groups) input
relation for an INTERSECT. Implement that, and upgrade prepunion's estimation
of the number of rows returned by setops so that there's some amount of sanity
in the estimate of which one is smaller.
This completes my project of improving usage of hashing for duplicate
elimination (aggregate functions with DISTINCT remain undone, but that's
for some other day).
As with the previous patches, this means we can INTERSECT/EXCEPT on datatypes
that can hash but not sort, and it means that INTERSECT/EXCEPT without ORDER
BY are no longer certain to produce sorted output.
as per my recent proposal:
1. Fold SortClause and GroupClause into a single node type SortGroupClause.
We were already relying on them to be struct-equivalent, so using two node
tags wasn't accomplishing much except to get in the way of comparing items
with equal().
2. Add an "eqop" field to SortGroupClause to carry the associated equality
operator. This is cheap for the parser to get at the same time it's looking
up the sort operator, and storing it eliminates the need for repeated
not-so-cheap lookups during planning. In future this will also let us
represent GROUP/DISTINCT operations on datatypes that have hash opclasses
but no btree opclasses (ie, they have equality but no natural sort order).
The previous representation simply didn't work for that, since its only
indicator of comparison semantics was a sort operator.
3. Add a hasDistinctOn boolean to struct Query to explicitly record whether
the distinctClause came from DISTINCT or DISTINCT ON. This allows removing
some complicated and not 100% bulletproof code that attempted to figure
that out from the distinctClause alone.
This patch doesn't in itself create any new capability, but it's necessary
infrastructure for future attempts to use hash-based grouping for DISTINCT
and UNION/INTERSECT/EXCEPT.
taking the maximum of any child rel's width, we should weight the widths
proportionally to the number of rows expected from each child. In hindsight
this is obviously correct because row width is really a proxy for the total
physical size of the relation. Per discussion with Scott Carey (bug #4264).
we had several code paths where a physical tlist could be used for the input
to a Sort node, which is a dumb idea because any unneeded table columns will
increase the volume of data the sort has to push around.
(Unfortunately the easy-looking fix of calling disuse_physical_tlist during
make_sort_xxx doesn't work because in most cases we're already committed to
the current input tlist --- it's been marked with sort column numbers, or
we've built grouping column numbers using it, etc. The tlist has to be
selected properly at the calling level before we start constructing sort-col
information. This is easy enough to do, we were just failing to take the
point into consideration.)
Back-patch to 8.3. I believe the problem probably exists clear back to 7.4
when the physical tlist optimization was added, but I'm afraid to back-patch
further than 8.3 without a great deal more study than I want to put into it.
The code in this area has drifted a lot over time. The real-world importance
of these code paths is uncertain anyway --- I think in many cases we'd
probably prefer hash-based methods.
no particular need to do get_op_opfamily_properties() while building an
indexscan plan. Postpone that lookup until executor start. This simplifies
createplan.c a lot more than it complicates nodeIndexscan.c, and makes things
more uniform since we already had to do it that way for RowCompare
expressions. Should be a bit faster too, at least for plans that aren't
re-used many times, since we avoid palloc'ing and perhaps copying the
intermediate list data structure.
instead of plan time. Extend the amgettuple API so that the index AM returns
a boolean indicating whether the indexquals need to be rechecked, and make
that rechecking happen in nodeIndexscan.c (currently the only place where
it's expected to be needed; other callers of index_getnext are just erroring
out for now). For the moment, GIN and GIST have stub logic that just always
sets the recheck flag to TRUE --- I'm hoping to get Teodor to handle pushing
that control down to the opclass consistent() functions. The planner no
longer pays any attention to amopreqcheck, and that catalog column will go
away in due course.
predictable manner; in particular that if you say ORDER BY output-column-ref,
it will in fact sort by that specific column even if there are multiple
syntactic matches. An example is
SELECT random() AS a, random() AS b FROM ... ORDER BY b, a;
While the use-case for this might be a bit debatable, it worked as expected
in earlier releases, so we should preserve the behavior for 8.3. Per my
recent proposal.
While at it, fix convert_subquery_pathkeys() to handle RelabelType stripping
in both directions; it needs this for the same reasons make_sort_from_pathkeys
does.
to be able to discard top-level RelabelType nodes on *both* sides of the
equivalence-class-to-target-list comparison, since make_pathkey_from_sortinfo
might either add or remove a RelabelType. Also fix the latter to do the
removal case cleanly. Per example from Peter.
RelabelType nodes when the sort key is binary-compatible with the sort
operator rather than having exactly its input type. We did this correctly
for index columns but not sort keys, leading to failure to notice that
a varchar index matches an ORDER BY request. This requires a bit more work
in make_sort_from_pathkeys, but not anyplace else that I can find.
Per bug report and subsequent discussion.
is using mark/restore but not rewind or backward-scan capability. Insert a
materialize plan node between a mergejoin and its inner child if the inner
child is a sort that is expected to spill to disk. The materialize shields
the sort from the need to do mark/restore and thereby allows it to perform
its final merge pass on-the-fly; while the materialize itself is normally
cheap since it won't spill to disk unless the number of tuples with equal
key values exceeds work_mem.
Greg Stark, with some kibitzing from Tom Lane.
need be returned. We keep a heap of the current best N tuples and sift-up
new tuples into it as we scan the input. For M input tuples this means
only about M*log(N) comparisons instead of M*log(M), not to mention a lot
less workspace when N is small --- avoiding spill-to-disk for large M
is actually the most attractive thing about it. Patch includes planner
and executor support for invoking this facility in ORDER BY ... LIMIT
queries. Greg Stark, with some editorialization by moi.
are mostly excluded by constraints: do the CE test a bit earlier to save
some adjust_appendrel_attrs() work on excluded children, and arrange to
use array indexing rather than rt_fetch() to fetch RTEs in the main body
of the planner. The latter is something I'd wanted to do for awhile anyway,
but seeing list_nth_cell() as 35% of the runtime gets one's attention.
is still needed despite cleanups in setrefs.c, because the point is to
let the inserted Result node compute a different tlist than its input
node does. Per example from Jeremy Drake.
useless substructure for its RangeTblEntry nodes. (I chose to keep using the
same struct node type and just zero out the link fields for unneeded info,
rather than making a separate ExecRangeTblEntry type --- it seemed too
fragile to have two different rangetable representations.)
Along the way, put subplans into a list in the toplevel PlannedStmt node,
and have SubPlan nodes refer to them by list index instead of direct pointers.
Vadim wanted to do that years ago, but I never understood what he was on about
until now. It makes things a *whole* lot more robust, because we can stop
worrying about duplicate processing of subplans during expression tree
traversals. That's been a constant source of bugs, and it's finally gone.
There are some consequent simplifications yet to be made, like not using
a separate EState for subplans in the executor, but I'll tackle that later.
plan nodes, so that the executor does not need to get these items from
the range table at runtime. This will avoid needing to include these
fields in the compact range table I'm expecting to make the executor use.
Hashing for aggregation purposes still needs work, so it's not time to
mark any cross-type operators as hashable for general use, but these cases
work if the operators are so marked by hand in the system catalogs.
columns procost and prorows, to allow simple user adjustment of the estimated
cost of a function call, as well as control of the estimated number of rows
returned by a set-returning function. We might eventually wish to extend this
to allow function-specific estimation routines, but there seems to be
consensus that we should try a simple constant estimate first. In particular
this provides a relatively simple way to control the order in which different
WHERE clauses are applied in a plan node, which is a Good Thing in view of the
fact that the recent EquivalenceClass planner rewrite made that much less
predictable than before.
representation of equivalence classes of variables. This is an extensive
rewrite, but it brings a number of benefits:
* planner no longer fails in the presence of "incomplete" operator families
that don't offer operators for every possible combination of datatypes.
* avoid generating and then discarding redundant equality clauses.
* remove bogus assumption that derived equalities always use operators
named "=".
* mergejoins can work with a variety of sort orders (e.g., descending) now,
instead of tying each mergejoinable operator to exactly one sort order.
* better recognition of redundant sort columns.
* can make use of equalities appearing underneath an outer join.
which comparison operators to use for plan nodes involving tuple comparison
(Agg, Group, Unique, SetOp). Formerly the executor looked up the default
equality operator for the datatype, which was really pretty shaky, since it's
possible that the data being fed to the node is sorted according to some
nondefault operator class that could have an incompatible idea of equality.
The planner knows what it has sorted by and therefore can provide the right
equality operator to use. Also, this change moves a couple of catalog lookups
out of the executor and into the planner, which should help startup time for
pre-planned queries by some small amount. Modify the planner to remove some
other cavalier assumptions about always being able to use the default
operators. Also add "nulls first/last" info to the Plan node for a mergejoin
--- neither the executor nor the planner can cope yet, but at least the API is
in place.
per-column options for btree indexes. The planner's support for this is still
pretty rudimentary; it does not yet know how to plan mergejoins with
nondefault ordering options. The documentation is pretty rudimentary, too.
I'll work on improving that stuff later.
Note incompatible change from prior behavior: ORDER BY ... USING will now be
rejected if the operator is not a less-than or greater-than member of some
btree opclass. This prevents less-than-sane behavior if an operator that
doesn't actually define a proper sort ordering is selected.
cases. Operator classes now exist within "operator families". While most
families are equivalent to a single class, related classes can be grouped
into one family to represent the fact that they are semantically compatible.
Cross-type operators are now naturally adjunct parts of a family, without
having to wedge them into a particular opclass as we had done originally.
This commit restructures the catalogs and cleans up enough of the fallout so
that everything still works at least as well as before, but most of the work
needed to actually improve the planner's behavior will come later. Also,
there are not yet CREATE/DROP/ALTER OPERATOR FAMILY commands; the only way
to create a new family right now is to allow CREATE OPERATOR CLASS to make
one by default. I owe some more documentation work, too. But that can all
be done in smaller pieces once this infrastructure is in place.
(e.g. "INSERT ... VALUES (...), (...), ...") and elsewhere as allowed
by the spec. (e.g. similar to a FROM clause subselect). initdb required.
Joe Conway and Tom Lane.
clauses containing no variables and no volatile functions. Such a clause
can be used as a one-time qual in a gating Result plan node, to suppress
plan execution entirely when it is false. Even when the clause is true,
putting it in a gating node wins by avoiding repeated evaluation of the
clause. In previous PG releases, query_planner() would do this for
pseudoconstant clauses appearing at the top level of the jointree, but
there was no ability to generate a gating Result deeper in the plan tree.
To fix it, get rid of the special case in query_planner(), and instead
process pseudoconstant clauses through the normal RestrictInfo qual
distribution mechanism. When a pseudoconstant clause is found attached to
a path node in create_plan(), pull it out and generate a gating Result at
that point. This requires special-casing pseudoconstants in selectivity
estimation and cost_qual_eval, but on the whole it's pretty clean.
It probably even makes the planner a bit faster than before for the normal
case of no pseudoconstants, since removing pull_constant_clauses saves one
useless traversal of the qual tree. Per gripe from Phil Frost.
the partial index predicate in the scan's "recheck condition". Otherwise,
if the scan becomes lossy for lack of bitmap memory, we would fail to enforce
that returned rows satisfy the predicate. Noted while studying bug #2441
from Arjen van der Meijden.
support both FOR UPDATE and FOR SHARE in one command, as well as both
NOWAIT and normal WAIT behavior. The more general code is actually
simpler and cleaner.
implied by the predicate of a partial index being used to scan a table.
However, this optimization is unsafe in an UPDATE, DELETE, or SELECT FOR
UPDATE query, because the quals need to be rechecked by EvalPlanQual if
there's an update conflict. Per example from Jean-Samuel Reynaud.
to avoid sharing substructure with the lower-level indexquals. This is
currently only an issue if there are SubPlans in the indexquals, which is
uncommon but not impossible --- see bug #2218 reported by Nicholas Vinen.
We use the same kluge for indexqual vs indexqualorig in the index scans
themselves ... would be nice to clean this up someday.
"ctid IN (list)" will still work after we convert IN to ScalarArrayOpExpr.
Make some minor efficiency improvements while at it, such as ensuring that
multiple TIDs are fetched in physical heap order. And fix EXPLAIN so that
it shows what's really going on for a TID scan.
qualification when the underlying operator is indexable and useOr is true.
That is, indexkey op ANY (ARRAY[...]) is effectively translated into an
OR combination of one indexscan for each array element. This only works
for bitmap index scans, of course, since regular indexscans no longer
support OR'ing of scans. There are still some loose ends to clean up
before changing 'x IN (list)' to translate as a ScalarArrayOpExpr;
for instance predtest.c ought to be taught about it. But this gets the
basic functionality in place.
comment line where output as too long, and update typedefs for /lib
directory. Also fix case where identifiers were used as variable names
in the backend, but as typedefs in ecpg (favor the backend for
indenting).
Backpatch to 8.1.X.
optimization for subquery and function scan nodes: we can't just do it
unconditionally, we still have to check whether there is any need for
a whole-row Var. I had been thinking that these node types couldn't
have any system columns, which is true, but that loop is also checking
for attno zero, ie, whole-row Var. Fix comment to not be so misleading.
Per test case from Richard Huxton.
make_restrictinfo_from_bitmapqual. The likelihood of finding duplicates
seems much less than in the AND-subclause case, and the cost much higher,
because OR lists with hundreds or even thousands of subclauses are not
uncommon. Per discussion with Ilia Kantor and andrew@supernews.
predicate_implied_by() to detect redundant filter conditions, but forgot
that predicate_implied_by() assumes its first argument contains only
immutable functions. Add a check to guarantee that. Also, test to see
if filter conditions can be discarded because they are redundant with
the predicate of a partial index.
or OFFSET clauses by using estimate_expression_value(). The main advantage
of this is that if the expression is a Param and we have a value for the
Param, we'll use that value rather than defaulting. Also, fix some
thinkos in the logic for combining LIMIT/OFFSET with an externally
supplied tuple fraction (this covers cases like EXISTS(...LIMIT...)).
And make sure the results of all this are shown by EXPLAIN. Per a
gripe from Merlin Moncure.
planning logic for bitmap indexscans. Partial indexes create corner
cases in which a scan might be done with no explicit index qual conditions,
and the code wasn't handling those cases nicely. Also be a little
tenser about eliminating redundant clauses in the generated plan.
Per report from Dmitry Karasik.
output targetlist of the Unique or HashAgg plan. This code was OK when
written, but subsequent changes to use "physical tlists" where possible
had broken it: given an input subplan that has extra variables added to
avoid a projection step, it would copy those extra variables into the
upper tlist, which is pointless since a projection has to happen anyway.
propagated inside an outer join. In particular, given
LEFT JOIN ON (A = B) WHERE A = constant, we cannot conclude that
B = constant at the top level (B might be null instead), but we
can nonetheless put a restriction B = constant into the quals for
B's relation, since no inner-side rows not meeting that condition
can contribute to the final result. Similarly, given
FULL JOIN USING (J) WHERE J = constant, we can't directly conclude
that either input J variable = constant, but it's OK to push such
quals into each input rel. Per recent gripe from Kim Bisgaard.
Along the way, remove 'valid_everywhere' flag from RestrictInfo,
as on closer analysis it was not being used for anything, and was
defined backwards anyway.
a new PlannerInfo struct, which is passed around instead of the bare
Query in all the planning code. This commit is essentially just a
code-beautification exercise, but it does open the door to making
larger changes to the planner data structures without having to muck
with the widely-known Query struct.
aren't doing anything useful (ie, neither selection nor projection).
Also, extend to SubqueryScan the hacks already in place to avoid
unnecessary ExecProject calls when the result would just be the same
tuple the subquery already delivered. This saves some overhead in
UNION and other set operations, as well as avoiding overhead for
unflatten-able subqueries. Per example from Sokolov Yura.
or bitmap), use pred_test to be a little smarter about cases where a
filter clause is logically unnecessary. This may be overkill for the
plain indexscan case, but it's definitely useful for OR'd bitmap scans.
node, as this behavior is now better done as a bitmap OR indexscan.
This allows considerable simplification in nodeIndexscan.c itself as
well as several planner modules concerned with indexscan plan generation.
Also we can improve the sharing of code between regular and bitmap
indexscans, since they are now working with nigh-identical Plan nodes.
but the code is basically working. Along the way, rewrite the entire
approach to processing OR index conditions, and make it work in join
cases for the first time ever. orindxpath.c is now basically obsolete,
but I left it in for the time being to allow easy comparison testing
against the old implementation.
logic operations during planning. Seems cleaner to create two new Path
node types, instead --- this avoids duplication of cost-estimation code.
Also, create an enable_bitmapscan GUC parameter to control use of bitmap
plans.
scans, using in-memory tuple ID bitmaps as the intermediary. The planner
frontend (path creation and cost estimation) is not there yet, so none
of this code can be executed. I have tested it using some hacked planner
code that is far too ugly to see the light of day, however. Committing
now so that the bulk of the infrastructure changes go in before the tree
drifts under me.
few palloc's. I also chose to eliminate the restype and restypmod fields
entirely, since they are redundant with information stored in the node's
contained expression; re-examining the expression at need seems simpler
and more reliable than trying to keep restype/restypmod up to date.
initdb forced due to change in contents of stored rules.
structs. There are many places in the planner where we were passing
both a rel and an index to subroutines, and now need only pass the
index struct. Notationally simpler, and perhaps a tad faster.
Formerly, if such a clause contained no aggregate functions we mistakenly
treated it as equivalent to WHERE. Per spec it must cause the query to
be treated as a grouped query of a single group, the same as appearance
of aggregate functions would do. Also, the HAVING filter must execute
after aggregate function computation even if it itself contains no
aggregate functions.
Also performed an initial run through of upgrading our Copyright date to
extend to 2005 ... first run here was very simple ... change everything
where: grep 1996-2004 && the word 'Copyright' ... scanned through the
generated list with 'less' first, and after, to make sure that I only
picked up the right entries ...
In the past, we used a 'Lispy' linked list implementation: a "list" was
merely a pointer to the head node of the list. The problem with that
design is that it makes lappend() and length() linear time. This patch
fixes that problem (and others) by maintaining a count of the list
length and a pointer to the tail node along with each head node pointer.
A "list" is now a pointer to a structure containing some meta-data
about the list; the head and tail pointers in that structure refer
to ListCell structures that maintain the actual linked list of nodes.
The function names of the list API have also been changed to, I hope,
be more logically consistent. By default, the old function names are
still available; they will be disabled-by-default once the rest of
the tree has been updated to use the new API names.
equivalent sort expressions to use was broken: you can't just look
at the relation membership, you have to actually grovel over the
individual Vars in each expression. I think this did work when it
was written, but it was broken by subsequent optimizations that made
join relations not propagate every single input variable upward.
Must find the Var that got propagated, not choose one at random.
Per bug report from Daniel O'Neill.
check instead of hardwiring assumptions that only certain plan node types
can appear at the places where we are testing. This was always a pretty
fragile assumption, and it turns out to be broken in 7.4 for certain cases
involving IN-subselect tests that need type coercion.
Also, modify code that builds finished Plan tree so that node types that
don't do projection always copy their input node's targetlist, rather than
having the tlist passed in from the caller. The old method makes it too
easy to write broken code that thinks it can modify the tlist when it
cannot.
pointer type when it is not necessary to do so.
For future reference, casting NULL to a pointer type is only necessary
when (a) invoking a function AND either (b) the function has no prototype
OR (c) the function is a varargs function.
regular qpqual ('filter condition'), add special-purpose code to
nodeIndexscan.c to recheck them. This ends being almost no net addition
of code, because the removal of planner code balances out the extra
executor code, but it is significantly more efficient when a lossy
operator is involved in an OR indexscan. The old implementation had
to recheck the entire indexqual in such cases.
with index qual clauses in the Path representation. This saves a little
work during createplan and (probably more importantly) allows reuse of
cached selectivity estimates during indexscan planning. Also fix latent
bug: wrong plan would have been generated for a 'special operator' used
in a nestloop-inner-indexscan join qual, because the special operator
would not have gotten into the list of quals to recheck. This bug is
only latent because at present the special-operator code could never
trigger on a join qual, but sooner or later someone will want to do it.
join conditions in which each OR subclause includes a constraint on
the same relation. This implements the other useful side-effect of
conversion to CNF format, without its unpleasant side-effects. As
per pghackers discussion of a few weeks ago.
the hashclauses field of the parent HashJoin. This avoids problems with
duplicated links to SubPlans in hash clauses, as per report from
Andrew Holm-Hansen.
pghackers proposal of 8-Nov. All the existing cross-type comparison
operators (int2/int4/int8 and float4/float8) have appropriate support.
The original proposal of storing the right-hand-side datatype as part of
the primary key for pg_amop and pg_amproc got modified a bit in the event;
it is easier to store zero as the 'default' case and only store a nonzero
when the operator is actually cross-type. Along the way, remove the
long-since-defunct bigbox_ops operator class.
Remove the 'strategy map' code, which was a large amount of mechanism
that no longer had any use except reverse-mapping from procedure OID to
strategy number. Passing the strategy number to the index AM in the
first place is simpler and faster.
This is a preliminary step in planned support for cross-datatype index
operations. I'm committing it now since the ScanKeyEntryInitialize()
API change touches quite a lot of files, and I want to commit those
changes before the tree drifts under me.
datatype by array_eq and array_cmp; use this to solve problems with memory
leaks in array indexing support. The parser's equality_oper and ordering_oper
routines also use the cache. Change the operator search algorithms to look
for appropriate btree or hash index opclasses, instead of assuming operators
named '<' or '=' have the right semantics. (ORDER BY ASC/DESC now also look
at opclasses, instead of assuming '<' and '>' are the right things.) Add
several more index opclasses so that there is no regression in functionality
for base datatypes. initdb forced due to catalog additions.
yet, though). Avoid using nth() to fetch tlist entries; provide a
common routine get_tle_by_resno() to search a tlist for a particular
resno. This replaces a couple uses of nth() and a dozen hand-coded
search loops. Also, replace a few uses of nth(length-1, list) with
llast().
subplan it starts with, as they may be needed at upper join levels.
See comments added to code for the non-obvious reason why. Per bug report
from Robert Creager.
node emits only those vars that are actually needed above it in the
plan tree. (There were comments in the code suggesting that this was
done at some point in the dim past, but for a long time we have just
made join nodes emit everything that either input emitted.) Aside from
being marginally more efficient, this fixes the problem noted by Peter
Eisentraut where a join above an IN-implemented-as-join might fail,
because the subplan targetlist constructed in the latter case didn't
meet the expectation of including everything.
Along the way, fix some places that were O(N^2) in the targetlist
length. This is not all the trouble spots for wide queries by any
means, but it's a step forward.
silently resolving them to type TEXT. This is comparable to what we
do when faced with UNKNOWN in CASE, UNION, and other contexts. It gets
rid of this and related annoyances:
select distinct f1, '' from int4_tbl;
ERROR: Unable to identify an ordering operator '<' for type unknown
This was discussed many moons ago, but no one got round to fixing it.
some cases of redundant clauses that were formerly not caught. We have
to special-case this because the clauses involved never get attached to
the same join restrictlist and so the existing logic does not notice
that they are redundant.
of an index can now be a computed expression instead of a simple variable.
Restrictions on expressions are the same as for predicates (only immutable
functions, no sub-selects). This fixes problems recently introduced with
inlining SQL functions, because the inlining transformation is applied to
both expression trees so the planner can still match them up. Along the
way, improve efficiency of handling index predicates (both predicates and
index expressions are now cached by the relcache) and fix 7.3 oversight
that didn't record dependencies of predicate expressions.
dropped. The simplest fix for INSERT/UPDATE cases turns out to be for
preptlist.c to insert NULLs of a known-good type (I used INT4) rather
than making them match the deleted column's type. Since the representation
of NULL is actually datatype-independent, this should work fine.
I also re-reverted the patch to disable the use_physical_tlist optimization
in the presence of dropped columns. It still doesn't look worth the
trouble to be smarter, if there are no other bugs to fix.
Added a regression test to catch future problems in this area.
where the table contains dropped columns. If the columns are dropped,
then their types may be gone as well, which causes ExecTypeFromTL() to
fail if the dropped columns appear in a plan node's tlist. This could
be worked around but I don't think the optimization is valuable enough
to be worth the trouble.
the column by table OID and column number, if it's a simple column
reference. Along the way, get rid of reskey/reskeyop fields in Resdoms.
Turns out that representation was not convenient for either the planner
or the executor; we can make the planner deliver exactly what the
executor wants with no more effort.
initdb forced due to change in stored rule representation.
utility statement (DeclareCursorStmt) with a SELECT query dangling from
it, rather than a SELECT query with a few unusual fields in it. Add
code to determine whether a planned query can safely be run backwards.
If DECLARE CURSOR specifies SCROLL, ensure that the plan can be run
backwards by adding a Materialize plan node if it can't. Without SCROLL,
you get an error if you try to fetch backwards from a cursor that can't
handle it. (There is still some discussion about what the exact
behavior should be, but this is necessary infrastructure in any case.)
Along the way, make EXPLAIN DECLARE CURSOR work.
in the case where the node immediately above the scan is a Hash, Sort,
or Material node. In these cases it's better to do the projection
so that we don't store unneeded columns in the hash/sort/materialize
table. Per discussion a few days ago with Anagh Lal.
rid of the assumption that sizeof(Oid)==sizeof(int). This is one small
step towards someday supporting 8-byte OIDs. For the moment, it doesn't
do much except get rid of a lot of unsightly casts.
nodes where it's not really necessary. In many cases where the scan node
is not the topmost plan node (eg, joins, aggregation), it's possible to
just return the table tuple directly instead of generating an intermediate
projection tuple. In preliminary testing, this reduced the CPU time
needed for 'SELECT COUNT(*) FROM foo' by about 10%.
There are two implementation techniques: the executor understands a new
JOIN_IN jointype, which emits at most one matching row per left-hand row,
or the result of the IN's sub-select can be fed through a DISTINCT filter
and then joined as an ordinary relation.
Along the way, some minor code cleanup in the optimizer; notably, break
out most of the jointree-rearrangement preprocessing in planner.c and
put it in a new file prep/prepjointree.c.
join_references(), it's practical to consolidate all join_references()
processing into the set_plan_references traversal in setrefs.c. This
seems considerably cleaner than the old way where we did it for join
quals in createplan.c and for targetlists in setrefs.c.
containing a volatile function), rather than only on 'Var = Var' clauses
as before. This makes it practical to do flatten_join_alias_vars at the
start of planning, which in turn eliminates a bunch of klugery inside the
planner to deal with alias vars. As a free side effect, we now detect
implied equality of non-Var expressions; for example in
SELECT ... WHERE a.x = b.y and b.y = 42
we will deduce a.x = 42 and use that as a restriction qual on a. Also,
we can remove the restriction introduced 12/5/02 to prevent pullup of
subqueries whose targetlists contain sublinks.
Still TODO: make statistical estimation routines in selfuncs.c and costsize.c
smarter about expressions that are more complex than plain Vars. The need
for this is considerably greater now that we have to be able to estimate
the suitability of merge and hash join techniques on such expressions.
HAVING quals. Normally this is an insignificant effect --- but it
will not be insignificant when these clauses contain sub-selects.
The added costs cannot affect the planning of the query containing
them, but they might have an impact when the query is a sub-query
of a larger one.
so that all executable expression nodes inherit from a common supertype
Expr. This is somewhat of an exercise in code purity rather than any
real functional advance, but getting rid of the extra Oper or Func node
formerly used in each operator or function call should provide at least
a little space and speed improvement.
initdb forced by changes in stored-rules representation.
to plan nodes, not vice-versa. All executor state nodes now inherit from
struct PlanState. Copying of plan trees has been simplified by not
storing a list of SubPlans in Plan nodes (eliminating duplicate links).
The executor still needs such a list, but it can build it during
ExecutorStart since it has to scan the plan tree anyway.
No initdb forced since no stored-on-disk structures changed, but you
will need a full recompile because of node-numbering changes.
instead of only one. This should speed up planning (only one hash path
to consider for a given pair of relations) as well as allow more effective
hashing, when there are multiple hashable joinclauses.
parameter to allow it to be forced off for comparison purposes.
Add ORDER BY clauses to a bunch of regression test queries that will
otherwise produce randomly-ordered output in the new regime.
of groups produced by GROUP BY. This improves the accuracy of planning
estimates for grouped subselects, and is needed to check whether a
hashed aggregation plan risks memory overflow.
node now does its own grouping of the input rows, and has no need for a
preceding GROUP node in the plan pipeline. This allows elimination of
the misnamed tuplePerGroup option for GROUP, and actually saves more code
in nodeGroup.c than it costs in nodeAgg.c, as well as being presumably
faster. Restructure the API of query_planner so that we do not commit to
using a sorted or unsorted plan in query_planner; instead grouping_planner
makes the decision. (Right now it isn't any smarter than query_planner
was, but that will change as soon as it has the option to select a hash-
based aggregation step.) Despite all the hackery, no initdb needed since
only in-memory node types changed.
to be flexible about assignment casts without introducing ambiguity in
operator/function resolution. Introduce a well-defined promotion hierarchy
for numeric datatypes (int2->int4->int8->numeric->float4->float8).
Change make_const to initially label numeric literals as int4, int8, or
numeric (never float8 anymore).
Explicitly mark Func and RelabelType nodes to indicate whether they came
from a function call, explicit cast, or implicit cast; use this to do
reverse-listing more accurately and without so many heuristics.
Explicit casts to char, varchar, bit, varbit will truncate or pad without
raising an error (the pre-7.2 behavior), while assigning to a column without
any explicit cast will still raise an error for wrong-length data like 7.3.
This more nearly follows the SQL spec than 7.2 behavior (we should be
reporting a 'completion condition' in the explicit-cast cases, but we have
no mechanism for that, so just do silent truncation).
Fix some problems with enforcement of typmod for array elements;
it didn't work at all in 'UPDATE ... SET array[n] = foo', for example.
Provide a generalized array_length_coerce() function to replace the
specialized per-array-type functions that used to be needed (and were
missing for NUMERIC as well as all the datetime types).
Add missing conversions int8<->float4, text<->numeric, oid<->int8.
initdb forced.
rather than a Query node; this allows set_plan_references to recurse
into subplans correctly. Fixes core dump on full outer joins in
subplans. Also, invoke preprocess_expression on function RTEs'
function expressions. This seems to fix the planner's problems with
outer-level Vars in function RTEs.
some kibitzing from Tom Lane. Not everything works yet, and there's
no documentation or regression test, but let's commit this so Joe
doesn't need to cope with tracking changes in so many files ...
lists to join RTEs, attach a list of Vars and COALESCE expressions that will
replace the join's alias variables during planning. This simplifies
flatten_join_alias_vars while still making it easy to fix up varno references
when transforming the query tree. Add regression test cases for interactions
of subqueries with outer joins.
now has an RTE of its own, and references to its outputs now are Vars
referencing the JOIN RTE, rather than CASE-expressions. This allows
reverse-listing in ruleutils.c to use the correct alias easily, rather
than painfully reverse-engineering the alias namespace as it used to do.
Also, nested FULL JOINs work correctly, because the result of the inner
joins are simple Vars that the planner can cope with. This fixes a bug
reported a couple times now, notably by Tatsuo on 18-Nov-01. The alias
Vars are expanded into COALESCE expressions where needed at the very end
of planning, rather than during parsing.
Also, beginnings of support for showing plan qualifier expressions in
EXPLAIN. There are probably still cases that need work.
initdb forced due to change of stored-rule representation.
pgsql-hackers. pg_opclass now has a row for each opclass supported by each
index AM, not a row for each opclass name. This allows pg_opclass to show
directly whether an AM supports an opclass, and furthermore makes it possible
to store additional information about an opclass that might be AM-dependent.
pg_opclass and pg_amop now store "lossy" and "haskeytype" information that we
previously expected the user to remember to provide in CREATE INDEX commands.
Lossiness is no longer an index-level property, but is associated with the
use of a particular operator in a particular index opclass.
Along the way, IndexSupportInitialize now uses the syscaches to retrieve
pg_amop and pg_amproc entries. I find this reduces backend launch time by
about ten percent, at the cost of a couple more special cases in catcache.c's
IndexScanOK.
Initial work by Oleg Bartunov and Teodor Sigaev, further hacking by Tom Lane.
initdb forced.
of costsize.c routines to pass Query root, so that costsize can figure
more things out by itself and not be so dependent on its callers to tell
it everything it needs to know. Use selectivity of hash or merge clause
to estimate number of tuples processed internally in these joins
(this is more useful than it would've been before, since eqjoinsel is
somewhat more accurate than before).
create_index_paths are not immediately discarded, but are available for
subsequent planner work. This allows avoiding redundant syscache lookups
in several places. Change interface to operator selectivity estimation
procedures to allow faster and more flexible estimation.
Initdb forced due to change of pg_proc entries for selectivity functions!
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.
maintained for each cache entry. A cache entry will not be freed until
the matching ReleaseSysCache call has been executed. This eliminates
worries about cache entries getting dropped while still in use. See
my posting to pg-hackers of even date for more info.
joins, and clean things up a good deal at the same time. Append plan node
no longer hacks on rangetable at runtime --- instead, all child tables are
given their own RT entries during planning. Concept of multiple target
tables pushed up into execMain, replacing bug-prone implementation within
nodeAppend. Planner now supports generating Append plans for inheritance
sets either at the top of the plan (the old way) or at the bottom. Expanding
at the bottom is appropriate for tables used as sources, since they may
appear inside an outer join; but we must still expand at the top when the
target of an UPDATE or DELETE is an inheritance set, because we actually need
a different targetlist and junkfilter for each target table in that case.
Fortunately a target table can't be inside an outer join... Bizarre mutual
recursion between union_planner and prepunion.c is gone --- in fact,
union_planner doesn't really have much to do with union queries anymore,
so I renamed it grouping_planner.
ExecutorRun. This allows LIMIT to work in a view. Also, LIMIT in a
cursor declaration will behave in a reasonable fashion, whereas before
it was overridden by the FETCH count.
SQL92 semantics, including support for ALL option. All three can be used
in subqueries and views. DISTINCT and ORDER BY work now in views, too.
This rewrite fixes many problems with cross-datatype UNIONs and INSERT/SELECT
where the SELECT yields different datatypes than the INSERT needs. I did
that by making UNION subqueries and SELECT in INSERT be treated like
subselects-in-FROM, thereby allowing an extra level of targetlist where the
datatype conversions can be inserted safely.
INITDB NEEDED!
(Don't forget that an alias is required.) Views reimplemented as expanding
to subselect-in-FROM. Grouping, aggregates, DISTINCT in views actually
work now (he says optimistically). No UNION support in subselects/views
yet, but I have some ideas about that. Rule-related permissions checking
moved out of rewriter and into executor.
INITDB REQUIRED!
right thing with variable-free clauses that contain noncachable functions,
such as 'WHERE random() < 0.5' --- these are evaluated once per
potential output tuple. Expressions that contain only Params are
now candidates to be indexscan quals --- for example, 'var = ($1 + 1)'
can now be indexed. Cope with RelabelType nodes atop potential indexscan
variables --- this oversight prevents 7.0.* from recognizing some
potentially indexscanable situations.
memory contexts. Currently, only leaks in expressions executed as
quals or projections are handled. Clean up some old dead cruft in
executor while at it --- unused fields in state nodes, that sort of thing.
materialized tupleset is small enough) instead of a temporary relation.
This was something I was thinking of doing anyway for performance, and Jan
says he needs it for TOAST because he doesn't want to cope with toasting
noname relations. With this change, the 'noname table' support in heap.c
is dead code, and I have accordingly removed it. Also clean up 'noname'
plan handling in planner --- nonames are either sort or materialize plans,
and it seems less confusing to handle them separately under those names.
think that both sides of indexqual look like index keys. An example is
create table inside (f1 float8 primary key);
create table outside (g1 float8, g2 float8);
select * from inside,outside where f1 = atan2(g1+1, g2);
ERROR: ExecInitIndexScan: both left and right ops are rel-vars
(note that failure is potentially platform-dependent). Solution is a
cleanup I had had in mind to make anyway: functional index keys should
be represented as Var nodes in the fixed indexqual, just like regular
index keys.
WHERE in a place where it can be part of a nestloop inner indexqual.
As the code stood, it put the same physical sub-Plan node into both
indxqual and indxqualorig of the IndexScan plan node. That confused
later processing in the optimizer (which expected that tracing the
subPlan list would visit each subplan node exactly once), and would
probably have blown up in the executor if the planner hadn't choked first.
Fix by making the 'fixed' indexqual be a complete deep copy of the
original indexqual, rather than trying to share nodes below the topmost
operator node. This had further ramifications though, because we were
making the aforesaid list of sub-Plan nodes during SS_process_sublinks
which is run before construction of the 'fixed' indexqual, meaning that
the copy of the sub-Plan didn't show up in that list. Fix by rearranging
logic so that the sub-Plan list is built by the final set_plan_references
pass, not in SS_process_sublinks. This may sound like a mess, but it's
actually a good deal cleaner now than it was before, because we are no
longer dependent on the assumption that planning will never make a copy
of a sub-Plan node.
costs using the inner path's parent->rows count as the number of tuples
processed per inner scan iteration. This is wrong when we are using an
inner indexscan with indexquals based on join clauses, because the rows
count in a Relation node reflects the selectivity of the restriction
clauses for that rel only. Upshot was that if join clause was very
selective, we'd drastically overestimate the true cost of the join.
Fix is to calculate correct output-rows estimate for an inner indexscan
when the IndexPath node is created and save it in the path node.
Change of path node doesn't require initdb, since path nodes don't
appear in saved rules.
accesses versus sequential accesses, a (very crude) estimate of the
effects of caching on random page accesses, and cost to evaluate WHERE-
clause expressions. Export critical parameters for this model as SET
variables. Also, create SET variables for the planner's enable flags
(enable_seqscan, enable_indexscan, etc) so that these can be controlled
more conveniently than via PGOPTIONS.
Planner now estimates both startup cost (cost before retrieving
first tuple) and total cost of each path, so it can optimize queries
with LIMIT on a reasonable basis by interpolating between these costs.
Same facility is a win for EXISTS(...) subqueries and some other cases.
Redesign pathkey representation to achieve a major speedup in planning
(I saw as much as 5X on a 10-way join); also minor changes in planner
to reduce memory consumption by recycling discarded Path nodes and
not constructing unnecessary lists.
Minor cleanups to display more-plausible costs in some cases in
EXPLAIN output.
Initdb forced by change in interface to index cost estimation
functions.
fields in JoinPaths --- turns out that we do need that after all :-(.
Also, rearrange planner so that only one RelOptInfo is created for a
particular set of joined base relations, no matter how many different
subsets of relations it can be created from. This saves memory and
processing time compared to the old method of making a bunch of RelOptInfos
and then removing the duplicates. Clean up the jointree iteration logic;
not sure if it's better, but I sure find it more readable and plausible
now, particularly for the case of 'bushy plans'.
SELECT DISTINCT ON (expr [, expr ...]) targetlist ...
and there is a check to make sure that the user didn't specify an ORDER BY
that's incompatible with the DISTINCT operation.
Reimplement nodeUnique and nodeGroup to use the proper datatype-specific
equality function for each column being compared --- they used to do
bitwise comparisons or convert the data to text strings and strcmp().
(To add insult to injury, they'd look up the conversion functions once
for each tuple...) Parse/plan representation of DISTINCT is now a list
of SortClause nodes.
initdb forced by querytree change...
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 ...
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.
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...
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.
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.
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.
> >
> > 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
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...
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.
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.
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.)
so remove them from MergeJoin node. Hack together a partial
solution for commuted mergejoin operators --- yesterday
a mergejoin int4 = int8 would crash if the planner decided to
commute it, today it works. The planner's representation of
mergejoins really needs a rewrite though.
Also, further testing of mergejoin ops in opr_sanity regress test.
Ok. I made patches replacing all of "#if FALSE" or "#if 0" to "#ifdef
NOT_USED" for current. I have tested these patches in that the
postgres binaries are identical.
of some global variables to support subselects and calls union_planner().
Calls to SS_replace_correlation_vars() and SS_process_sublinks() in
query_planner() before planning.
Get rid of #ifdef INDEXSCAN_PATCH in createplan.c.
Pass List* of Aggregs into executor, and create needed array there.
No longer need to double-processs Aggregs with second copy in Query.
Fix crash when doing:
select sum(x+1) from test where 1 > 0;
nestloop's join clauses doesn't work in some cases:
* 1. fix_indxqual_references may change varattno-s in
* inner_indxqual;
* 2. clauses may be commuted
Subject: [HACKERS] linux/alpha patches
These patches lay the groundwork for a Linux/Alpha port. The port doesn't
actually work unless you tweak the linker to put all the pointers in the
first 32 bits of the address space, but it's at least a start. It
implements the test-and-set instruction in Alpha assembly, and also fixes
a lot of pointer-to-integer conversions, which is probably good anyway.
The problem is that the function arguments are not considered as possible key
candidates for index scan and so only a sequential scan is possible inside
the body of a function. I have therefore made some patches to the optimizer
so that indices are now used also by functions. I have also moved the plan
debug message from pg_eval to pg_plan so that it is printed also for plans
genereated for function execution. I had also to add an index rescan to the
executor because it ignored the parameters set in the execution state, they
were flagged as runtime variables in ExecInitIndexScan but then never used
by the executor so that the scan were always done with any key=1. Very odd.
This means that an index rescan is now done twice for each function execution
which uses an index, the first time when the index scan is initialized and
the second when the actual function arguments are finally available for the
execution. I don't know what is the cost of an double index scan but I
suppose it is anyway less than the cost of a full sequential scan, at leat
for large tables. This is my patch, you must also add -DINDEXSCAN_PATCH in
Makefile.global to enable the changes.
Submitted by: Massimo Dal Zotto <dz@cs.unitn.it>