Add PlaceHolderVar wrappers as needed to make UNION ALL sub-select output
expressions appear non-constant and distinct from each other. This makes
the world safe for add_child_rel_equivalences to do what it does. Before,
it was possible for that function to add identical expressions to different
EquivalenceClasses, which logically should imply merging such ECs, which
would be wrong; or to improperly add a constant to an EquivalenceClass,
drastically changing its behavior. Per report from Teodor Sigaev.
The only currently known consequence of this bug is "MergeAppend child's
targetlist doesn't match MergeAppend" planner failures in 9.1 and later.
I am suspicious that there may be other failure modes that could affect
older release branches; but in the absence of any hard evidence, I'll
refrain from back-patching further than 9.1.
If we use a PlaceHolderVar from the outer relation in an inner indexscan,
we need to reference the PlaceHolderVar as such as the value to be passed
in from the outer relation. The previous code effectively tried to
reconstruct the PHV from its component expression, which doesn't work since
(a) the Vars therein aren't necessarily bubbled up far enough, and (b) it
would be the wrong semantics anyway because of the possibility that the PHV
is supposed to have gone to null at some point before the current join.
Point (a) led to "variable not found in subplan target list" planner
errors, but point (b) would have led to silently wrong answers.
Per report from Roger Niederland.
If the right-hand side of a semijoin is unique, then we can treat it like a
normal join (or another way to say that is: we don't need to explicitly
unique-ify the data before doing it as a normal join). We were recognizing
such cases when the RHS was a sub-query with appropriate DISTINCT or GROUP
BY decoration, but there's another way: if the RHS is a plain relation with
unique indexes, we can check if any of the indexes prove the output is
unique. Most of the infrastructure for that was there already in the join
removal code, though I had to rearrange it a bit. Per reflection about a
recent example in pgsql-performance.
Add a column pg_class.relallvisible to remember the number of pages that
were all-visible according to the visibility map as of the last VACUUM
(or ANALYZE, or some other operations that update pg_class.relpages).
Use relallvisible/relpages, instead of an arbitrary constant, to estimate
how many heap page fetches can be avoided during an index-only scan.
This is pretty primitive and will no doubt see refinements once we've
acquired more field experience with the index-only scan mechanism, but
it's way better than using a constant.
Note: I had to adjust an underspecified query in the window.sql regression
test, because it was changing answers when the plan changed to use an
index-only scan. Some of the adjacent tests perhaps should be adjusted
as well, but I didn't do that here.
When a btree index contains all columns required by the query, and the
visibility map shows that all tuples on a target heap page are
visible-to-all, we don't need to fetch that heap page. This patch depends
on the previous patches that made the visibility map reliable.
There's a fair amount left to do here, notably trying to figure out a less
chintzy way of estimating the cost of an index-only scan, but the core
functionality seems ready to commit.
Robert Haas and Ibrar Ahmed, with some previous work by Heikki Linnakangas.
The constraint exclusion feature checks for contradictions among scan
restriction clauses, as well as contradictions between those clauses and a
table's CHECK constraints. The first aspect of this testing can be useful
for non-table relations (such as subqueries or functions-in-FROM), but the
feature was coded with only the CHECK case in mind so we were applying it
only to plain-table RTEs. Move the relation_excluded_by_constraints call
so that it is applied to all RTEs not just plain tables. With the default
setting of constraint_exclusion this results in no extra work, but with
constraint_exclusion = ON we will detect optimizations that we missed
before (at the cost of more planner cycles than we expended before).
Per a gripe from Gunnlaugur Þór Briem. Experimentation with
his example also showed we were not being very bright about the case where
constraint exclusion is proven within a subquery within UNION ALL, so tweak
the code to allow set_append_rel_pathlist to recognize such cases.
Formerly, set_subquery_pathlist and other creators of plans for subqueries
saved only the rangetable and rowMarks lists from the lower-level
PlannerInfo. But there's no reason not to remember the whole PlannerInfo,
and indeed this turns out to simplify matters in a number of places.
The immediate reason for doing this was so that the subroot will still be
accessible when we're trying to extract column statistics out of an
already-planned subquery. But now that I've done it, it seems like a good
code-beautification effort in its own right.
I also chose to get rid of the transient subrtable and subrowmark fields in
SubqueryScan nodes, in favor of having setrefs.c look up the subquery's
RelOptInfo. That required changing all the APIs in setrefs.c to pass
PlannerInfo not PlannerGlobal, which was a large but quite mechanical
transformation.
One side-effect not foreseen at the beginning is that this finally broke
inheritance_planner's assumption that replanning the same subquery RTE N
times would necessarily give interchangeable results each time. That
assumption was always pretty risky, but now we really have to make a
separate RTE for each instance so that there's a place to carry the
separate subroots.
A PlaceHolderVar's expression might contain another, lower-level
PlaceHolderVar. If the outer PlaceHolderVar is used, the inner one
certainly will be also, and so we have to make sure that both of them get
into the placeholder_list with correct ph_may_need values during the
initial pre-scan of the query (before deconstruct_jointree starts).
We did this correctly for PlaceHolderVars appearing in the query quals,
but overlooked the issue for those appearing in the top-level targetlist;
with the result that nested placeholders referenced only in the targetlist
did not work correctly, as illustrated in bug #6154.
While at it, add some error checking to find_placeholder_info to ensure
that we don't try to create new placeholders after it's too late to do so;
they have to all be created before deconstruct_jointree starts.
Back-patch to 8.4 where the PlaceHolderVar mechanism was introduced.
glibc renders random() thread-safe by wrapping a futex lock around it;
testing reveals that this limits the performance of pgbench on machines
with many CPU cores. Rather than switching to random_r(), which is
only available on GNU systems and crashes unless you use undocumented
alchemy to initialize the random state properly, switch to our built-in
implementation of erand48(), which is both thread-safe and concurrent.
Since the list of reasons not to use the operating system's erand48()
is getting rather long, rename ours to pg_erand48() (and similarly
for our implementations of lrand48() and srand48()) and just always
use those. We were already doing this on Cygwin anyway, and the
glibc implementation is not quite thread-safe, so pgbench wouldn't
be able to use that either.
Per discussion with Tom Lane.
Regular aggregate functions in combination with, or within the arguments
of, window functions are OK per spec; they have the semantics that the
aggregate output rows are computed and then we run the window functions
over that row set. (Thus, this combination is not really useful unless
there's a GROUP BY so that more than one aggregate output row is possible.)
The case without GROUP BY could fail, as recently reported by Jeff Davis,
because sloppy construction of the Agg node's targetlist resulted in extra
references to possibly-ungrouped Vars appearing outside the aggregate
function calls themselves. See the added regression test case for an
example.
Fixing this requires modifying the API of flatten_tlist and its underlying
function pull_var_clause. I chose to make pull_var_clause's API for
aggregates identical to what it was already doing for placeholders, since
the useful behaviors turn out to be the same (error, report node as-is, or
recurse into it). I also tightened the error checking in this area a bit:
if it was ever valid to see an uplevel Var, Aggref, or PlaceHolderVar here,
that was a long time ago, so complain instead of ignoring them.
Backpatch into 9.1. The failure exists in 8.4 and 9.0 as well, but seeing
that it only occurs in a basically-useless corner case, it doesn't seem
worth the risks of changing a function API in a minor release. There might
be third-party code using pull_var_clause.
The previous coding failed to account properly for the costs of evaluating
the input expressions of aggregates and window functions, as seen in a
recent gripe from Claudio Freire. (I said at the time that it wasn't
counting these costs at all; but on closer inspection, it was effectively
charging these costs once per output tuple. That is completely wrong for
aggregates, and not exactly right for window functions either.)
There was also a hard-wired assumption that aggregates and window functions
had procost 1.0, which is now fixed to respect the actual cataloged costs.
The costing of WindowAgg is still pretty bogus, since it doesn't try to
estimate the effects of spilling data to disk, but that seems like a
separate issue.
This area was a few bricks shy of a load, and badly under-commented too.
We have to ensure that the generated targetlist entries for a set-operation
node expose the correct collation for each entry, since higher-level
processing expects the tlist to reflect the true ordering of the plan's
output.
This hackery wouldn't be necessary if SortGroupClause carried collation
info ... but making it do so would inject more pain in the parser than
would be saved here. Still, we might want to rethink that sometime.
Instead of playing cute games with pathkeys, just build a direct
representation of the intended sub-select, and feed it through
query_planner to get a Path for the index access. This is a bit slower
than 9.1's previous method, since we'll duplicate most of the overhead of
query_planner; but since the whole optimization only applies to rather
simple single-table queries, that probably won't be much of a problem in
practice. The advantage is that we get to do the right thing when there's
a partial index that needs the implicit IS NOT NULL clause to be usable.
Also, although this makes planagg.c be a bit more closely tied to the
ordering of operations in grouping_planner, we can get rid of some coupling
to lower-level parts of the planner. Per complaint from Marti Raudsepp.
All expression nodes now have an explicit output-collation field, unless
they are known to only return a noncollatable data type (such as boolean
or record). Also, nodes that can invoke collation-aware functions store
a separate field that is the collation value to pass to the function.
This avoids confusion that arises when a function has collatable inputs
and noncollatable output type, or vice versa.
Also, replace the parser's on-the-fly collation assignment method with
a post-pass over the completed expression tree. This allows us to use
a more complex (and hopefully more nearly spec-compliant) assignment
rule without paying for it in extra storage in every expression node.
Fix assorted bugs in the planner's handling of collations by making
collation one of the defining properties of an EquivalenceClass and
by converting CollateExprs into discardable RelabelType nodes during
expression preprocessing.
This patch implements data-modifying WITH queries according to the
semantics that the updates all happen with the same command counter value,
and in an unspecified order. Therefore one WITH clause can't see the
effects of another, nor can the outer query see the effects other than
through the RETURNING values. And attempts to do conflicting updates will
have unpredictable results. We'll need to document all that.
This commit just fixes the code; documentation updates are waiting on
author.
Marko Tiikkaja and Hitoshi Harada
This commit provides the core code and documentation needed. A contrib
module test case will follow shortly.
Shigeru Hanada, Jan Urbanski, Heikki Linnakangas
This adds collation support for columns and domains, a COLLATE clause
to override it per expression, and B-tree index support.
Peter Eisentraut
reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
This is a heavily revised version of builtin_knngist_core-0.9. The
ordering operators are no longer mixed in with actual quals, which would
have confused not only humans but significant parts of the planner.
Instead, ordering operators are carried separately throughout planning and
execution.
Since the API for ambeginscan and amrescan functions had to be changed
anyway, this commit takes the opportunity to rationalize that a bit.
RelationGetIndexScan no longer forces a premature index_rescan call;
instead, callers of index_beginscan must call index_rescan too. Aside from
making the AM-side initialization logic a bit less peculiar, this has the
advantage that we do not make a useless extra am_rescan call when there are
runtime key values. AMs formerly could not assume that the key values
passed to amrescan were actually valid; now they can.
Teodor Sigaev and Tom Lane
As per the ancient comment for set_rel_width, it really wasn't much good
for relations that aren't plain tables: it would never find any stats and
would always fall back on datatype-based estimates, which are often pretty
silly. Fix that by copying up width estimates from the subquery planning
process.
At some point we might want to do this for CTEs too, but that would be a
significantly more invasive patch because the sub-PlannerInfo is no longer
accessible by the time it's needed. I refrained from doing anything about
that, partly for fear of breaking the unmerged CTE-related patches.
In passing, also generate less bogus width estimates for whole-row Vars.
Per a gripe from Jon Nelson.
Formerly, we could convert a UNION ALL structure inside a subquery-in-FROM
into an appendrel, as a side effect of pulling up the subquery into its
parent; but top-level UNION ALL always caused use of plan_set_operations().
That didn't matter too much because you got an Append-based plan either
way. However, now that the appendrel code can do things with MergeAppend,
it's worthwhile to hack up the top-level case so it also uses appendrels.
This is a bit of a stopgap; but going much further than this will require
a major rewrite of the planner's set-operations support, which I'm not
prepared to undertake now. For the moment let's grab the low-hanging fruit.
Per my recent proposal, get rid of all the direct inspection of indexes
and manual generation of paths in planagg.c. Instead, set up
EquivalenceClasses for the aggregate argument expressions, and let the
regular path generation logic deal with creating paths that can satisfy
those sort orders. This makes planagg.c a bit more visible to the rest
of the planner than it was originally, but the approach is basically a lot
cleaner than before. A major advantage of doing it this way is that we get
MIN/MAX optimization on inheritance trees (using MergeAppend of indexscans)
practically for free, whereas in the old way we'd have had to add a whole
lot more duplicative logic.
One small disadvantage of this approach is that MIN/MAX aggregates can no
longer exploit partial indexes having an "x IS NOT NULL" predicate, unless
that restriction or something that implies it is specified in the query.
The previous implementation was able to use the added "x IS NOT NULL"
condition as an extra predicate proof condition, but in this version we
rely entirely on indexes that are considered usable by the main planning
process. That seems a fair tradeoff for the simplicity and functionality
gained.
Zoltan Boszormenyi exhibited a test case in which planning time was
dominated by construction of EquivalenceClasses and PathKeys that had no
actual relevance to the query (and in fact got discarded immediately).
This happened because we generated PathKeys describing the sort ordering of
every index on every table in the query, and only after that checked to see
if the sort ordering was relevant. The EC/PK construction code is O(N^2)
in the number of ECs, which is all right for the intended number of such
objects, but it gets out of hand if there are ECs for lots of irrelevant
indexes.
To fix, twiddle the handling of mergeclauses a little bit to ensure that
every interesting EC is created before we begin path generation. (This
doesn't cost anything --- in fact I think it's a bit cheaper than before
--- since we always eventually created those ECs anyway.) Then, if an
index column can't be found in any pre-existing EC, we know that that sort
ordering is irrelevant for the query. Instead of creating a useless EC,
we can just not build a pathkey for the index column in the first place.
The index will still be considered if it's useful for non-order-related
reasons, but we will think of its output as unsorted.
This patch eliminates the former need to sort the output of an Append scan
when an ordered scan of an inheritance tree is wanted. This should be
particularly useful for fast-start cases such as queries with LIMIT.
Original patch by Greg Stark, with further hacking by Hans-Jurgen Schonig,
Robert Haas, and Tom Lane.
This patch merges the responsibility for NOT-flattening into
eval_const_expressions' processing. It wasn't done that way originally
because prepqual.c is far older than eval_const_expressions. But putting
this work into eval_const_expressions saves one pass over the qual trees,
and in fact saves even more than that because we can exploit the knowledge
that the subexpressions have already been recursively simplified. Doing it
this way also lets us do it uniformly over all expressions, whereas
prepqual.c formerly just did it at top level to save cycles. That should
improve the planner's ability to recognize logically-equivalent constructs.
While at it, also add the ability to fold a NOT into BooleanTest and
NullTest constructs (the latter only for the scalar-datatype case).
Per discussion of bug #5702.
The point of a PlaceHolderVar is to allow a non-strict expression to be
evaluated below an outer join, after which its value bubbles up like a Var
and can be forced to NULL when the outer join's semantics require that.
However, there was a serious design oversight in that, namely that we
didn't ensure that there was actually a correct place in the plan tree
to evaluate the placeholder :-(. It may be necessary to delay evaluation
of an outer join to ensure that a placeholder that should be evaluated
below the join can be evaluated there. Per recent bug report from Kirill
Simonov.
Back-patch to 8.4 where the PlaceHolderVar mechanism was introduced.
In these cases a qual can get marked with the removable rel in its
required_relids, but this is just to schedule its evaluation correctly, not
because it really depends on the rel. We were assuming that, in effect,
we could throw away *all* quals so marked, which is nonsense. Tighten up
the logic to be a little more paranoid about which quals belong to the
outer join being considered for removal, and arrange for all quals that
don't belong to be updated so they will still get evaluated correctly.
Also fix another problem that happened to be exposed by this test case,
which was that make_join_rel() was failing to notice some cases where
a constant-false qual could be used to prove a join relation empty. If it's
a pushed-down constant false, then the relation is empty even if it's an
outer join, because the qual applies after the outer join expansion.
Per report from Nathan Grange. Back-patch into 9.0.
relation using the general PARAM_EXEC executor parameter mechanism, rather
than the ad-hoc kluge of passing the outer tuple down through ExecReScan.
The previous method was hard to understand and could never be extended to
handle parameters coming from multiple join levels. This patch doesn't
change the set of possible plans nor have any significant performance effect,
but it's necessary infrastructure for future generalization of the concept
of an inner indexscan plan.
ExecReScan's second parameter is now unused, so it's removed.
The logic for determining whether to materialize has been significantly
overhauled for 9.0. In case there should be any doubt about whether
materialization is a win in any particular case, this should provide a
convenient way of seeing what happens without it; but even with enable_material
turned off, we still materialize in cases where it is required for
correctness.
Thanks to Tom Lane for the review.
This patch allows the frame to start from CURRENT ROW (in either RANGE or
ROWS mode), and it also adds support for ROWS n PRECEDING and ROWS n FOLLOWING
start and end points. (RANGE value PRECEDING/FOLLOWING isn't there yet ---
the grammar works, but that's all.)
Hitoshi Harada, reviewed by Pavel Stehule
parse analysis phase, rather than at execution time. This makes parameter
handling work the same as it does in ordinary plannable queries, and in
particular fixes the incompatibility that Pavel pointed out with plpgsql's
new handling of variable references. plancache.c gets a little bit
grottier, but the alternatives seem worse.
to be just a minor extension of the previous patch that made "x IS NULL"
indexable, because we can treat the IS NOT NULL condition as if it were
"x < NULL" or "x > NULL" (depending on the index's NULLS FIRST/LAST option),
just like IS NULL is treated like "x = NULL". Aside from any possible
usefulness in its own right, this is an important improvement for
index-optimized MAX/MIN aggregates: it is now reliably possible to get
a column's min or max value cheaply, even when there are a lot of nulls
cluttering the interesting end of the index.
non-kluge method for controlling the order in which values are fed to an
aggregate function. At the same time eliminate the old implementation
restriction that DISTINCT was only supported for single-argument aggregates.
Possibly release-notable behavioral change: formerly, agg(DISTINCT x)
dropped null values of x unconditionally. Now, it does so only if the
agg transition function is strict; otherwise nulls are treated as DISTINCT
normally would, ie, you get one copy.
Andrew Gierth, reviewed by Hitoshi Harada
by adding a requirement that build_join_rel add new join RelOptInfos to the
appropriate list immediately at creation. Per report from Robert Haas,
the list_concat_unique_ptr() calls that this change eliminates were taking
the lion's share of the runtime in larger join problems. This doesn't do
anything to fix the fundamental combinatorial explosion in large join
problems, but it should push out the threshold of pain a bit further.
Note: because this changes the order in which joinrel lists are built,
it might result in changes in selected plans in cases where different
alternatives have exactly the same costs. There is one example in the
regression tests.
mergejoin to shield it from doing mark/restore and refetches. Put an explicit
flag in MergePath so we can centralize the logic that knows about this,
and add costing logic that considers using Materialize even when it's not
forced by the previously-existing considerations. This is in response to
a discussion back in August that suggested that materializing an inner
indexscan can be helpful when the refetch percentage is high enough.
a lot of strange behaviors that occurred in join cases. We now identify the
"current" row for every joined relation in UPDATE, DELETE, and SELECT FOR
UPDATE/SHARE queries. If an EvalPlanQual recheck is necessary, we jam the
appropriate row into each scan node in the rechecking plan, forcing it to emit
only that one row. The former behavior could rescan the whole of each joined
relation for each recheck, which was terrible for performance, and what's much
worse could result in duplicated output tuples.
Also, the original implementation of EvalPlanQual could not re-use the recheck
execution tree --- it had to go through a full executor init and shutdown for
every row to be tested. To avoid this overhead, I've associated a special
runtime Param with each LockRows or ModifyTable plan node, and arranged to
make every scan node below such a node depend on that Param. Thus, by
signaling a change in that Param, the EPQ machinery can just rescan the
already-built test plan.
This patch also adds a prohibition on set-returning functions in the
targetlist of SELECT FOR UPDATE/SHARE. This is needed to avoid the
duplicate-output-tuple problem. It seems fairly reasonable since the
other restrictions on SELECT FOR UPDATE are meant to ensure that there
is a unique correspondence between source tuples and result tuples,
which an output SRF destroys as much as anything else does.
execMain.c and into a new plan node type LockRows. Like the recent change
to put table updating into a ModifyTable plan node, this increases planning
flexibility by allowing the operations to occur below the top level of the
plan tree. It's necessary in any case to restore the previous behavior of
having FOR UPDATE locking occur before ModifyTable does.
This partially refactors EvalPlanQual to allow multiple rows-under-test
to be inserted into the EPQ machinery before starting an EPQ test query.
That isn't sufficient to fix EPQ's general bogosity in the face of plans
that return multiple rows per test row, though. Since this patch is
mostly about getting some plan node infrastructure in place and not about
fixing ten-year-old bugs, I will leave EPQ improvements for another day.
Another behavioral change that we could now think about is doing FOR UPDATE
before LIMIT, but that too seems like it should be treated as a followon
patch.
They are now handled by a new plan node type called ModifyTable, which is
placed at the top of the plan tree. In itself this change doesn't do much,
except perhaps make the handling of RETURNING lists and inherited UPDATEs a
tad less klugy. But it is necessary preparation for the intended extension of
allowing RETURNING queries inside WITH.
Marko Tiikkaja
is unique and is not referenced above the join. In this case the inner
side doesn't affect the query result and can be thrown away entirely.
Although perhaps nobody would ever write such a thing by hand, it's
a reasonably common case in machine-generated SQL.
The current implementation only recognizes the case where the inner side
is a simple relation with a unique index matching the query conditions.
This is enough for the use-cases that have been shown so far, but we
might want to try to handle other cases later.
Robert Haas, somewhat rewritten by Tom
an explicit model of rescan costs being different from first-time costs.
The costing of Material nodes in particular now has some visible relationship
to the actual runtime behavior, where before it was essentially fantasy.
This also fixes up a couple of places where different materialized plan types
were treated differently for no very good reason (probably just oversights).
A couple of the regression tests are affected, because the planner now chooses
to put the other relation on the inside of a nestloop-with-materialize.
So far as I can see both changes are sane, and the planner is now more
consistently following the expectation that it should prefer to materialize
the smaller of two relations.
Per a recent discussion with Robert Haas.
random number seed each time. This is how it used to work years ago, but
we got rid of the seed reset because it was resetting the main random()
sequence and thus having undesirable effects on the rest of the system.
To fix, establish a private random number state for each execution of
geqo(), and initialize the state using the new GUC variable geqo_seed.
People who want to experiment with different random searches can do so
by changing geqo_seed, but you'll always get the same plan for the same
value of geqo_seed (if holding all other planner inputs constant, of course).
The new state is kept in PlannerInfo by adding a "void *" field reserved
for use by join_search hooks. Most of the rather bulky code changes in
this commit are just arranging to pass PlannerInfo around to all the GEQO
functions (many of which formerly didn't receive it).
Andres Freund, with some editorialization by Tom
This alters various incidental uses of C++ key words to use other similar
identifiers, so that a C++ compiler won't choke outright. You still
(probably) need extern "C" { }; around the inclusion of backend headers.
based on a patch by Kurt Harriman <harriman@acm.org>
Also add a script cpluspluscheck to check for C++ compatibility in the
future. As of right now, this passes without error for me.
substituting a child rel's output expressions into the appendrel's restriction
clauses yields a pseudoconstant restriction. We might be able to skip scanning
that child rel entirely (if we get constant FALSE), or generate a one-time
filter. 8.3 more or less accidentally generated plans that weren't completely
stupid in these cases, but that was only because an extra recursive level of
subquery_planner() always occurred and allowed const-simplification to happen.
8.4's ability to pull up appendrel members with non-Var outputs exposes the
fact that we need to work harder here. Per gripe from Sergey Burladyan.
find_inheritance_children() and find_all_inheritors(). I got annoyed that
these are buried inside the planner but mostly used elsewhere. So, create
a new file catalog/pg_inherits.c and put them there, along with a couple
of other functions that search pg_inherits.
The code that modifies pg_inherits is (still) in tablecmds.c --- it's
kind of entangled with unrelated code that modifies pg_depend and other
stuff, so pulling it out seemed like a bigger change than I wanted to make
right now. But this file provides a natural home for it if anyone ever
gets around to that.
This commit just moves code around; it doesn't change anything, except
I succumbed to the temptation to make a couple of trivial optimizations
in typeInheritsFrom().
joins a bit better, ie, understand the differing cost functions for matched
and unmatched outer tuples. There is more that could be done in cost_hashjoin
but this already helps a great deal. Per discussions with Robert Haas.
PlaceHolderVar nodes in join quals appearing in or below the lowest
outer join that could null the subquery being pulled up. This improves
the planner's ability to recognize constant join quals, and probably
helps with detection of common sort keys (equivalence classes) as well.
Stefan Kaltenbrunner. The most reasonable behavior (at least for the near
term) seems to be to ignore the PlaceHolderVar and examine its argument
instead. In support of this, change the API of pull_var_clause() to allow
callers to request recursion into PlaceHolderVars. Currently
estimate_num_groups() is the only customer for that behavior, but where
there's one there may be others.
constants through full joins, as in
select * from tenk1 a full join tenk1 b using (unique1)
where unique1 = 42;
which should generate a fairly cheap plan where we apply the constraint
unique1 = 42 in each relation scan. This had been broken by my patch of
2008-06-27, which is now reverted in favor of a more invasive but hopefully
less incorrect approach. That patch was meant to prevent incorrect extraction
of OR'd indexclauses from OR conditions above an outer join. To do that
correctly we need more information than the outerjoin_delay flag can provide,
so add a nullable_relids field to RestrictInfo that records exactly which
relations are nulled by outer joins that are underneath a particular qual
clause. A side benefit is that we can make the test in create_or_index_quals
more specific: it is now smart enough to extract an OR'd indexclause into the
outer side of an outer join, even though it must not do so in the inner side.
The old coding couldn't distinguish these cases so it could not do either.
making pull_up_sublinks() construct a full-blown JoinExpr tree representation
of IN/EXISTS SubLinks that it is able to convert to semi or anti joins.
This makes pull_up_sublinks() a shade more complex, but the gain in semantic
clarity is worth it. I still have more to do in this area to address the
previously-discussed problems, but this commit in itself fixes at least one
bug in HEAD, as shown by added regression test case.
that are set up for execution with ExecPrepareExpr rather than going through
the full planner process. By introducing an explicit notion of "expression
planning", this patch also lays a bit of groundwork for maybe someday
allowing sub-selects in standalone expressions.
the default. This setting enables constraint exclusion checks only for
appendrel members (ie, inheritance children and UNION ALL arms), which are
the cases in which constraint exclusion is most likely to be useful. Avoiding
the overhead for simple queries that are unlikely to benefit should bring
the cost down to the point where this is a reasonable default setting.
Per today's discussion.
patch. This includes the ability to force the frame to cover the whole
partition, and the ability to make the frame end exactly on the current row
rather than its last ORDER BY peer. Supporting any more of the full SQL
frame-clause syntax will require nontrivial hacking on the window aggregate
code, so it'll have to wait for 8.5 or beyond.
we extended the appendrel mechanism to support UNION ALL optimization. The
reason nobody noticed was that we are not actually using attr_needed data for
appendrel children; hence it seems more reasonable to rip it out than fix it.
Back-patch to 8.2 because an Assert failure is possible in corner cases.
Per examination of an example from Jim Nasby.
In HEAD, also get rid of AppendRelInfo.col_mappings, which is quite inadequate
to represent UNION ALL situations; depend entirely on translated_vars instead.
that represent some expression that we desire to compute below the top level
of the plan, and then let that value "bubble up" as though it were a plain
Var (ie, a column value).
The immediate application is to allow sub-selects to be flattened even when
they are below an outer join and have non-nullable output expressions.
Formerly we couldn't flatten because such an expression wouldn't properly
go to NULL when evaluated above the outer join. Now, we wrap it in a
PlaceHolderVar and arrange for the actual evaluation to occur below the outer
join. When the resulting Var bubbles up through the join, it will be set to
NULL if necessary, yielding the correct results. This fixes a planner
limitation that's existed since 7.1.
In future we might want to use this mechanism to re-introduce some form of
Hellerstein's "expensive functions" optimization, ie place the evaluation of
an expensive function at the most suitable point in the plan tree.
implementation uses an in-memory hash table, so it will poop out for very
large recursive results ... but the performance characteristics of a
sort-based implementation would be pretty unpleasant too.
There are some unimplemented aspects: recursive queries must use UNION ALL
(should allow UNION too), and we don't have SEARCH or CYCLE clauses.
These might or might not get done for 8.4, but even without them it's a
pretty useful feature.
There are also a couple of small loose ends and definitional quibbles,
which I'll send a memo about to pgsql-hackers shortly. But let's land
the patch now so we can get on with other development.
Yoshiyuki Asaba, with lots of help from Tatsuo Ishii and Tom Lane
when user-defined functions used in a plan are modified. Also invalidate
plans when schemas, operators, or operator classes are modified; but for these
cases we just invalidate everything rather than tracking exact dependencies,
since these types of objects seldom change in a production database.
Tom Lane; loosely based on a patch by Martin Pihlak.
into nodes/nodeFuncs, so as to reduce wanton cross-subsystem #includes inside
the backend. There's probably more that should be done along this line,
but this is a start anyway.
subqueries into the same thing you'd have gotten from IN (except always with
unknownEqFalse = true, so as to get the proper semantics for an EXISTS).
I believe this fixes the last case within CVS HEAD in which an EXISTS could
give worse performance than an equivalent IN subquery.
The tricky part of this is that if the upper query probes the EXISTS for only
a few rows, the hashing implementation can actually be worse than the default,
and therefore we need to make a cost-based decision about which way to use.
But at the time when the planner generates plans for subqueries, it doesn't
really know how many times the subquery will be executed. The least invasive
solution seems to be to generate both plans and postpone the choice until
execution. Therefore, in a query that has been optimized this way, EXPLAIN
will show two subplans for the EXISTS, of which only one will actually get
executed.
There is a lot more that could be done based on this infrastructure: in
particular it's interesting to consider switching to the hash plan if we start
out using the non-hashed plan but find a lot more upper rows going by than we
expected. I have therefore left some minor inefficiencies in place, such as
initializing both subplans even though we will currently only use one.
level of a JOIN/ON clause, not only at top level of WHERE. (However, we
can't do this in an outer join's ON clause, unless the ANY/EXISTS refers
only to the nullable side of the outer join, so that it can effectively
be pushed down into the nullable side.) Per request from Kevin Grittner.
In passing, fix a bug in the initial implementation of EXISTS pullup:
it would Assert if the EXIST's WHERE clause used a join alias variable.
Since we haven't yet flattened join aliases when this transformation
happens, it's necessary to include join relids in the computed set of
RHS relids.
and anti joins. To do this, pass the SpecialJoinInfo struct for the current
join as an additional optional argument to operator join selectivity
estimation functions. This allows the estimator to tell not only what kind
of join is being formed, but which variable is on which side of the join;
a requirement long recognized but not dealt with till now. This also leaves
the door open for future improvements in the estimators, such as accounting
for the null-insertion effects of lower outer joins. I didn't do anything
about that in the current patch but the information is in principle deducible
from what's passed.
The patch also clarifies the definition of join selectivity for semi/anti
joins: it's the fraction of the left input that has (at least one) match
in the right input. This allows getting rid of some very fuzzy thinking
that I had committed in the original 7.4-era IN-optimization patch.
There's probably room to estimate this better than the present patch does,
but at least we know what to estimate.
Since I had to touch CREATE OPERATOR anyway to allow a variant signature
for join estimator functions, I took the opportunity to add a couple of
additional checks that were missing, per my recent message to -hackers:
* Check that estimator functions return float8;
* Require execute permission at the time of CREATE OPERATOR on the
operator's function as well as the estimator functions;
* Require ownership of any pre-existing operator that's modified by
the command.
I also moved the lookup of the functions out of OperatorCreate() and
into operatorcmds.c, since that seemed more consistent with most of
the other catalog object creation processes, eg CREATE TYPE.
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.
but seem like a separate patch since most of the remaining work is on the
executor side.) I took the opportunity to push selection of the grouping
operators for set operations into the parser where it belongs. Otherwise this
is just a small exercise in making prepunion.c consider both alternatives.
As with the recent DISTINCT patch, this means we can UNION on datatypes that
can hash but not sort, and it means that UNION without ORDER BY is 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.
the current query level that aren't in fact output parameters of the current
initPlans. (This means, for example, output parameters of regular subplans.)
To make this work correctly for output parameters coming from sibling
initplans requires rejiggering the API of SS_finalize_plan just a bit:
we need the siblings to be visible to it, rather than hidden as
SS_make_initplan_from_plan had been doing. This is really part of my response
to bug #4290, but I concluded this part probably shouldn't be back-patched,
since all that it's doing is to make a debugging cross-check tighter.
corresponding struct definitions. This allows other headers to avoid including
certain highly-loaded headers such as rel.h and relscan.h, instead using just
relcache.h, heapam.h or genam.h, which are more lightweight and thus cause less
unnecessary dependencies.
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.
eval_const_expressions needs to be passed the PlannerInfo ("root") structure,
because in some cases we want it to substitute values for Param nodes.
(So "constant" is not so constant as all that ...) This mistake partially
disabled optimization of unnamed extended-Query statements in 8.3: in
particular the LIKE-to-indexscan optimization would never be applied if the
LIKE pattern was passed as a parameter, and constraint exclusion depending
on a parameter value didn't work either.
are declared to return set, and consist of just a single SELECT. We
can replace the FROM-item with a sub-SELECT and then optimize much as
if we were dealing with a view. Patch from Richard Rowell, cleaned up
by me.
two buckets at the start, we create a number of buckets appropriate for the
estimated size of the table. This avoids a lot of expensive bucket-split
actions during initial index build on an already-populated table.
This is one of the two core ideas of Tom Raney and Shreya Bhargava's patch
to reduce hash index build time. I'm committing it separately to make it
easier for people to test the effects of this separately from the effects
of their other core idea (pre-sorting the index entries by bucket number).
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.
a relation as a reason to invalidate a plan when the relation changes. This
handles scenarios such as dropping/recreating a sequence that is referenced by
nextval('seq') in a cached plan. Rather than teach plancache.c all about
digging through plan trees to find regclass Consts, we charge the planner's
setrefs.c with making a list of the relation OIDs on which each plan depends.
That way the list can be built cheaply during a plan tree traversal that has
to happen anyway. Per bug #3662 and subsequent discussion.
eval_const_expressions simplifies this to just "WHERE false", but we have
already done pull_up_IN_clauses so the IN join will be done, or at least
planned, anyway. The trouble case comes when the sub-SELECT is itself a join
and we decide to implement the IN by unique-ifying the sub-SELECT outputs:
with no remaining reference to the output Vars in WHERE, we won't have
propagated the Vars up to the upper join point, leading to "variable not found
in subplan target lists" error. Fix by adding an extra scan of in_info_list
and forcing all Vars mentioned therein to be propagated up to the IN join
point. Per bug report from Miroslav Sulc.
join search order portion of the planner; this is specifically intended to
simplify developing a replacement for GEQO planning. Patch by Julius
Stroffek, editorialized on by me. I renamed make_one_rel_by_joins to
standard_join_search and make_rels_by_joins to join_search_one_level to better
reflect their place within this scheme.
(because they are uncorrelated with the immediate parent query). We were
charging the full run cost to the parent node, disregarding the fact that
only one row need be fetched for EXISTS. While this would only be a
cosmetic issue in most cases, it might possibly affect planning outcomes
if the parent query were itself a subquery to some upper query.
Per recent discussion with Steve Crawford.
columns, and the new version can be stored on the same heap page, we no longer
generate extra index entries for the new version. Instead, index searches
follow the HOT-chain links to ensure they find the correct tuple version.
In addition, this patch introduces the ability to "prune" dead tuples on a
per-page basis, without having to do a complete VACUUM pass to recover space.
VACUUM is still needed to clean up dead index entries, however.
Pavan Deolasee, with help from a bunch of other people.
and/or create plans for hypothetical situations; in particular, investigate
plans that would be generated using hypothetical indexes. This is a
heavily-rewritten version of the hooks proposed by Gurjeet Singh for his
Index Advisor project. In this formulation, the index advisor can be
entirely a loadable module instead of requiring a significant part to be
in the core backend, and plans can be generated for hypothetical indexes
without requiring the creation and rolling-back of system catalog entries.
The index advisor patch as-submitted is not compatible with these hooks,
but it needs significant work anyway due to other 8.2-to-8.3 planner
changes. With these hooks in the core backend, development of the advisor
can proceed as a pgfoundry project.
cheapest-startup-cost innerjoin indexscans, and make joinpath.c consider
both of these (when different) as the inside of a nestloop join. The
original design was based on the assumption that indexscan paths always
have negligible startup cost, and so total cost is the only important
figure of merit; an assumption that's obviously broken by bitmap
indexscans. This oversight could lead to choosing poor plans in cases
where fast-start behavior is more important than total cost, such as
LIMIT and IN queries. 8.1-vintage brain fade exposed by an example from
Chuck D.
