Since 1f39bce02, HashAgg nodes have had the ability to spill to disk when
memory consumption exceeds work_mem. That commit added new properties to
EXPLAIN ANALYZE to show the maximum memory usage and disk usage, however,
it didn't quite go as far as showing that information for parallel
workers. Since workers may have experienced something very different from
the main process, we should show this information per worker, as is done
in Sort.
Reviewed-by: Justin Pryzby
Reviewed-by: Jeff Davis
Discussion: https://postgr.es/m/CAApHDvpEKbfZa18mM1TD7qV6PG+w97pwCWq5tVD0dX7e11gRJw@mail.gmail.com
Backpatch-through: 13, where the hashagg spilling code was added.
Includes some manual cleanup of places that pgindent messed up,
most of which weren't per project style anyway.
Notably, it seems some people didn't absorb the style rules of
commit c9d297751, because there were a bunch of new occurrences
of function calls with a newline just after the left paren, all
with faulty expectations about how the rest of the call would get
indented.
Before discarding the old hash table in ExecReScanHashJoin, capture
its statistics, ensuring that we report the maximum hashtable size
across repeated rescans of the hash input relation. We can repurpose
the existing code for reporting hashtable size in parallel workers
to help with this, making the patch pretty small. This also ensures
that if rescans happen within parallel workers, we get the correct
maximums across all instances.
Konstantin Knizhnik and Tom Lane, per diagnosis by Thomas Munro
of a trouble report from Alvaro Herrera.
Discussion: https://postgr.es/m/20200323165059.GA24950@alvherre.pgsql
WITH TIES is an option to the FETCH FIRST N ROWS clause (the SQL
standard's spelling of LIMIT), where you additionally get rows that
compare equal to the last of those N rows by the columns in the
mandatory ORDER BY clause.
There was a proposal by Andrew Gierth to implement this functionality in
a more powerful way that would yield more features, but the other patch
had not been finished at this time, so we decided to use this one for
now in the spirit of incremental development.
Author: Surafel Temesgen <surafel3000@gmail.com>
Reviewed-by: Álvaro Herrera <alvherre@alvh.no-ip.org>
Reviewed-by: Tomas Vondra <tomas.vondra@2ndquadrant.com>
Discussion: https://postgr.es/m/CALAY4q9ky7rD_A4vf=FVQvCGngm3LOes-ky0J6euMrg=_Se+ag@mail.gmail.com
Discussion: https://postgr.es/m/87o8wvz253.fsf@news-spur.riddles.org.uk
Incremental Sort is an optimized variant of multikey sort for cases when
the input is already sorted by a prefix of the requested sort keys. For
example when the relation is already sorted by (key1, key2) and we need
to sort it by (key1, key2, key3) we can simply split the input rows into
groups having equal values in (key1, key2), and only sort/compare the
remaining column key3.
This has a number of benefits:
- Reduced memory consumption, because only a single group (determined by
values in the sorted prefix) needs to be kept in memory. This may also
eliminate the need to spill to disk.
- Lower startup cost, because Incremental Sort produce results after each
prefix group, which is beneficial for plans where startup cost matters
(like for example queries with LIMIT clause).
We consider both Sort and Incremental Sort, and decide based on costing.
The implemented algorithm operates in two different modes:
- Fetching a minimum number of tuples without check of equality on the
prefix keys, and sorting on all columns when safe.
- Fetching all tuples for a single prefix group and then sorting by
comparing only the remaining (non-prefix) keys.
We always start in the first mode, and employ a heuristic to switch into
the second mode if we believe it's beneficial - the goal is to minimize
the number of unnecessary comparions while keeping memory consumption
below work_mem.
This is a very old patch series. The idea was originally proposed by
Alexander Korotkov back in 2013, and then revived in 2017. In 2018 the
patch was taken over by James Coleman, who wrote and rewrote most of the
current code.
There were many reviewers/contributors since 2013 - I've done my best to
pick the most active ones, and listed them in this commit message.
Author: James Coleman, Alexander Korotkov
Reviewed-by: Tomas Vondra, Andreas Karlsson, Marti Raudsepp, Peter Geoghegan, Robert Haas, Thomas Munro, Antonin Houska, Andres Freund, Alexander Kuzmenkov
Discussion: https://postgr.es/m/CAPpHfdscOX5an71nHd8WSUH6GNOCf=V7wgDaTXdDd9=goN-gfA@mail.gmail.com
Discussion: https://postgr.es/m/CAPpHfds1waRZ=NOmueYq0sx1ZSCnt+5QJvizT8ndT2=etZEeAQ@mail.gmail.com
PostgreSQL provides set of template index access methods, where opclasses have
much freedom in the semantics of indexing. These index AMs are GiST, GIN,
SP-GiST and BRIN. There opclasses define representation of keys, operations on
them and supported search strategies. So, it's natural that opclasses may be
faced some tradeoffs, which require user-side decision. This commit implements
opclass parameters allowing users to set some values, which tell opclass how to
index the particular dataset.
This commit doesn't introduce new storage in system catalog. Instead it uses
pg_attribute.attoptions, which is used for table column storage options but
unused for index attributes.
In order to evade changing signature of each opclass support function, we
implement unified way to pass options to opclass support functions. Options
are set to fn_expr as the constant bytea expression. It's possible due to the
fact that opclass support functions are executed outside of expressions, so
fn_expr is unused for them.
This commit comes with some examples of opclass options usage. We parametrize
signature length in GiST. That applies to multiple opclasses: tsvector_ops,
gist__intbig_ops, gist_ltree_ops, gist__ltree_ops, gist_trgm_ops and
gist_hstore_ops. Also we parametrize maximum number of integer ranges for
gist__int_ops. However, the main future usage of this feature is expected
to be json, where users would be able to specify which way to index particular
json parts.
Catversion is bumped.
Discussion: https://postgr.es/m/d22c3a18-31c7-1879-fc11-4c1ce2f5e5af%40postgrespro.ru
Author: Nikita Glukhov, revised by me
Reviwed-by: Nikolay Shaplov, Robert Haas, Tom Lane, Tomas Vondra, Alvaro Herrera
While performing hash aggregation, track memory usage when adding new
groups to a hash table. If the memory usage exceeds work_mem, enter
"spill mode".
In spill mode, new groups are not created in the hash table(s), but
existing groups continue to be advanced if input tuples match. Tuples
that would cause a new group to be created are instead spilled to a
logical tape to be processed later.
The tuples are spilled in a partitioned fashion. When all tuples from
the outer plan are processed (either by advancing the group or
spilling the tuple), finalize and emit the groups from the hash
table. Then, create new batches of work from the spilled partitions,
and select one of the saved batches and process it (possibly spilling
recursively).
Author: Jeff Davis
Reviewed-by: Tomas Vondra, Adam Lee, Justin Pryzby, Taylor Vesely, Melanie Plageman
Discussion: https://postgr.es/m/507ac540ec7c20136364b5272acbcd4574aa76ef.camel@j-davis.com
When updating a table row with generated columns, only recompute those
generated columns whose base columns have changed in this update and
keep the rest unchanged. This can result in a significant performance
benefit. The required information was already kept in
RangeTblEntry.extraUpdatedCols; we just have to make use of it.
Reviewed-by: Pavel Stehule <pavel.stehule@gmail.com>
Discussion: https://www.postgresql.org/message-id/flat/b05e781a-fa16-6b52-6738-761181204567@2ndquadrant.com
It's already tracked via ExprState->parent, so we don't need to also
include it in ExprEvalStep. When that code originally was written
ExprState->parent didn't exist, but it since has been introduced in
6719b238e8.
Author: Andres Freund
Discussion: https://postgr.es/m/20191023163849.sosqbfs5yenocez3@alap3.anarazel.de
Commit 9b63c13f0 turns out to have been fundamentally misguided:
the parent node's subPlan list is by no means the only way in which
a child SubPlan node can be hooked into the outer execution state.
As shown in bug #16213 from Matt Jibson, we can also get short-lived
tuple table slots added to the outer es_tupleTable list. At this point
I have little faith that there aren't other possible connections as
well; the long time it took to notice this problem shows that this
isn't a heavily-exercised situation.
Therefore, revert that fix, returning to the coding that passed a
NULL parent plan pointer down to the transiently-built subexpressions.
That gives us a pretty good guarantee that they won't hook into the
outer executor state in any way. But then we need some other solution
to make SubPlans work. Adopt the solution speculated about in the
previous commit's log message: do expression initialization at plan
startup for just those VALUES rows containing SubPlans, abandoning the
goal of reclaiming memory intra-query for those rows. In practice it
seems unlikely that queries containing a vast number of VALUES rows
would be using SubPlans in them, so this should not give up much.
(BTW, this test case also refutes my claim in connection with the prior
commit that the issue only arises with use of LATERAL. That was just
wrong: some variants of SubLink always produce SubPlans.)
As with previous patch, back-patch to all supported branches.
Discussion: https://postgr.es/m/16213-871ac3bc208ecf23@postgresql.org
Previously, if the startup pruning logic proved that all child nodes
of an Append or MergeAppend could be pruned, we still kept one, just
to keep EXPLAIN from failing. The previous commit removed the
ruleutils.c limitation that required this kluge, so drop it. That
results in less-confusing EXPLAIN output, as per a complaint from
Yuzuko Hosoya.
David Rowley
Discussion: https://postgr.es/m/001001d4f44b$2a2cca50$7e865ef0$@lab.ntt.co.jp
This patch causes EXPLAIN to always assign a separate table alias to the
parent RTE of an append relation (inheritance set); before, such RTEs
were ignored if not actually scanned by the plan. Since the child RTEs
now always have that same alias to start with (cf. commit 55a1954da),
the net effect is that the parent RTE usually gets the alias used or
implied by the query text, and the children all get that alias with "_N"
appended. (The exception to "usually" is if there are duplicate aliases
in different subtrees of the original query; then some of those original
RTEs will also have "_N" appended.)
This results in more uniform output for partitioned-table plans than
we had before: the partitioned table itself gets the original alias,
and all child tables have aliases with "_N", rather than the previous
behavior where one of the children would get an alias without "_N".
The reason for giving the parent RTE an alias, even if it isn't scanned
by the plan, is that we now use the parent's alias to qualify Vars that
refer to an appendrel output column and appear above the Append or
MergeAppend that computes the appendrel. But below the append, Vars
refer to some one of the child relations, and are displayed that way.
This seems clearer than the old behavior where a Var that could carry
values from any child relation was displayed as if it referred to only
one of them.
While at it, change ruleutils.c so that the code paths used by EXPLAIN
deal in Plan trees not PlanState trees. This effectively reverts a
decision made in commit 1cc29fe7c, which seemed like a good idea at
the time to make ruleutils.c consistent with explain.c. However,
it's problematic because we'd really like to allow executor startup
pruning to remove all the children of an append node when possible,
leaving no child PlanState to resolve Vars against. (That's not done
here, but will be in the next patch.) This requires different handling
of subplans and initplans than before, but is otherwise a pretty
straightforward change.
Discussion: https://postgr.es/m/001001d4f44b$2a2cca50$7e865ef0$@lab.ntt.co.jp
In ad0bda5d24 I changed the EvalPlanQual machinery to store
substitution tuples in slot, instead of using plain HeapTuples. The
main motivation for that was that using HeapTuples will be inefficient
for future tableams. But it turns out that that conversion was buggy
for non-locking rowmarks - the wrong tuple descriptor was used to
create the slot.
As a secondary issue 5db6df0c0 changed ExecLockRows() to begin EPQ
earlier, to allow to fetch the locked rows directly into the EPQ
slots, instead of having to copy tuples around. Unfortunately, as Tom
complained, that forces some expensive initialization to happen
earlier.
As a third issue, the test coverage for EPQ was clearly insufficient.
Fixing the first issue is unfortunately not trivial: Non-locked row
marks were fetched at the start of EPQ, and we don't have the type
information for the rowmarks available at that point. While we could
change that, it's not easy. It might be worthwhile to change that at
some point, but to fix this bug, it seems better to delay fetching
non-locking rowmarks when they're actually needed, rather than
eagerly. They're referenced at most once, and in cases where EPQ
fails, might never be referenced. Fetching them when needed also
increases locality a bit.
To be able to fetch rowmarks during execution, rather than
initialization, we need to be able to access the active EPQState, as
that contains necessary data. To do so move EPQ related data from
EState to EPQState, and, only for EStates creates as part of EPQ,
reference the associated EPQState from EState.
To fix the second issue, change EPQ initialization to allow use of
EvalPlanQualSlot() to be used before EvalPlanQualBegin() (but
obviously still requiring EvalPlanQualInit() to have been done).
As these changes made struct EState harder to understand, e.g. by
adding multiple EStates, significantly reorder the members, and add a
lot more comments.
Also add a few more EPQ tests, including one that fails for the first
issue above. More is needed.
Reported-By: yi huang
Author: Andres Freund
Reviewed-By: Tom Lane
Discussion:
https://postgr.es/m/CAHU7rYZo_C4ULsAx_LAj8az9zqgrD8WDd4hTegDTMM1LMqrBsg@mail.gmail.comhttps://postgr.es/m/24530.1562686693@sss.pgh.pa.us
Backpatch: 12-, where the EPQ changes were introduced
Now that list_nth is O(1), there's no good reason to maintain a
separate array of RTE pointers rather than indexing into
estate->es_range_table. Deleting the array doesn't save all that
much either; but just on cleanliness grounds, it's better not to
have duplicate representations of the identical information.
Discussion: https://postgr.es/m/14960.1565384592@sss.pgh.pa.us
In 5f32b29c18 I changed the creation of HashState.hashkeys to
actually use HashState as the parent (instead of HashJoinState, which
was incorrect, as they were executed below HashState), to fix the
problem of hashkeys expressions otherwise relying on slot types
appropriate for HashJoinState, rather than HashState as would be
correct. That reliance was only introduced in 12, which is why it
previously worked to use HashJoinState as the parent (although I'd be
unsurprised if there were problematic cases).
Unfortunately that's not a sufficient solution, because before this
commit, the to-be-hashed expressions referenced inner/outer as
appropriate for the HashJoin, not Hash. That didn't have obvious bad
consequences, because the slots containing the tuples were put into
ecxt_innertuple when hashing a tuple for HashState (even though Hash
doesn't have an inner plan).
There are less common cases where this can cause visible problems
however (rather than just confusion when inspecting such executor
trees). E.g. "ERROR: bogus varno: 65000", when explaining queries
containing a HashJoin where the subsidiary Hash node's hash keys
reference a subplan. While normally hashkeys aren't displayed by
EXPLAIN, if one of those expressions references a subplan, that
subplan may be printed as part of the Hash node - which then failed
because an inner plan was referenced, and Hash doesn't have that.
It seems quite possible that there's other broken cases, too.
Fix the problem by properly splitting the expression for the HashJoin
and Hash nodes at plan time, and have them reference the proper
subsidiary node. While other workarounds are possible, fixing this
correctly seems easy enough. It was a pretty ugly hack to have
ExecInitHashJoin put the expression into the already initialized
HashState, in the first place.
I decided to not just split inner/outer hashkeys inside
make_hashjoin(), but also to separate out hashoperators and
hashcollations at plan time. Otherwise we would have ended up having
two very similar loops, one at plan time and the other during executor
startup. The work seems to more appropriately belong to plan time,
anyway.
Reported-By: Nikita Glukhov, Alexander Korotkov
Author: Andres Freund
Reviewed-By: Tom Lane, in an earlier version
Discussion: https://postgr.es/m/CAPpHfdvGVegF_TKKRiBrSmatJL2dR9uwFCuR+teQ_8tEXU8mxg@mail.gmail.com
Backpatch: 12-
This addresses a couple of issues in the code:
- Typos and inconsistencies in comments and function declarations.
- Removal of unreferenced function declarations.
- Removal of unnecessary compile flags.
- A cleanup error in regressplans.sh.
Author: Alexander Lakhin
Discussion: https://postgr.es/m/0c991fdf-2670-1997-c027-772a420c4604@gmail.com
We used the same slot to store a tuple from the index, and to store a
tuple from the table. That's not OK. It worked with the heap, because
heapam_getnextslot() stores a HeapTuple to the slot, and doesn't care how
large the tts_values/nulls arrays are. But when I played with a toy table
AM implementation that used a virtual tuple, it caused memory overruns.
In the passing, tidy up comments on the ioss_PscanLen fields.
This is still using the 2.0 version of pg_bsd_indent.
I thought it would be good to commit this separately,
so as to document the differences between 2.0 and 2.1 behavior.
Discussion: https://postgr.es/m/16296.1558103386@sss.pgh.pa.us
For amcanreorderby scans the nodeIndexscan.c's reorder queue holds
heap tuples, but the underlying table likely does not. Before this fix
we'd return different types of slots, depending on whether the tuple
came from the reorder queue, or from the index + table.
While that could be fixed by signalling that the node doesn't return a
fixed type of slot, it seems better to instead remove the separate
slot for the reorder queue, and use ExecForceStoreHeapTuple() to store
tuples from the queue. It's not particularly common to need
reordering, after all.
This reverts most of the iss_ReorderQueueSlot related changes to
nodeIndexscan.c made in 1a0586de36, except that now
ExecForceStoreHeapTuple() is used instead of ExecStoreHeapTuple().
Noticed when testing zheap against the in-core version of tableam.
Author: Andres Freund
This adds table_multi_insert(), and converts COPY FROM, the only user
of heap_multi_insert, to it.
A simple conversion of COPY FROM use slots would have yielded a
slowdown when inserting into a partitioned table for some
workloads. Different partitions might need different slots (both slot
types and their descriptors), and dropping / creating slots when
there's constant partition changes is measurable.
Thus instead revamp the COPY FROM buffering for partitioned tables to
allow to buffer inserts into multiple tables, flushing only when
limits are reached across all partition buffers. By only dropping
slots when there've been inserts into too many different partitions,
the aforementioned overhead is gone. By allowing larger batches, even
when there are frequent partition changes, we actuall speed such cases
up significantly.
By using slots COPY of very narrow rows into unlogged / temporary
might slow down very slightly (due to the indirect function calls).
Author: David Rowley, Andres Freund, Haribabu Kommi
Discussion:
https://postgr.es/m/20180703070645.wchpu5muyto5n647@alap3.anarazel.dehttps://postgr.es/m/20190327054923.t3epfuewxfqdt22e@alap3.anarazel.de
This moves bitmap heap scan support to below an optional tableam
callback. It's optional as the whole concept of bitmap heapscans is
fairly block specific.
This basically moves the work previously done in bitgetpage() into the
new scan_bitmap_next_block callback, and the direct poking into the
buffer done in BitmapHeapNext() into the new scan_bitmap_next_tuple()
callback.
The abstraction is currently somewhat leaky because
nodeBitmapHeapscan.c's prefetching and visibilitymap based logic
remains - it's likely that we'll later have to move more into the
AM. But it's not trivial to do so without introducing a significant
amount of code duplication between the AMs, so that's a project for
later.
Note that now nodeBitmapHeapscan.c and the associated node types are a
bit misnamed. But it's not clear whether renaming wouldn't be a cure
worse than the disease. Either way, that'd be best done in a separate
commit.
Author: Andres Freund
Reviewed-By: Robert Haas (in an older version)
Discussion: https://postgr.es/m/20180703070645.wchpu5muyto5n647@alap3.anarazel.de
This moves sample scan support to below tableam. It's not optional as
there is, in contrast to e.g. bitmap heap scans, no alternative way to
perform tablesample queries. If an AM can't deal with the block based
API, it will have to throw an ERROR.
The tableam callbacks for this are block based, but given the current
TsmRoutine interface, that seems to be required.
The new interface doesn't require TsmRoutines to perform visibility
checks anymore - that requires the TsmRoutine to know details about
the AM, which we want to avoid. To continue to allow taking the
returned number of tuples account SampleScanState now has a donetuples
field (which previously e.g. existed in SystemRowsSamplerData), which
is only incremented after the visibility check succeeds.
Author: Andres Freund
Discussion: https://postgr.es/m/20180703070645.wchpu5muyto5n647@alap3.anarazel.de
This is an SQL-standard feature that allows creating columns that are
computed from expressions rather than assigned, similar to a view or
materialized view but on a column basis.
This implements one kind of generated column: stored (computed on
write). Another kind, virtual (computed on read), is planned for the
future, and some room is left for it.
Reviewed-by: Michael Paquier <michael@paquier.xyz>
Reviewed-by: Pavel Stehule <pavel.stehule@gmail.com>
Discussion: https://www.postgresql.org/message-id/flat/b151f851-4019-bdb1-699e-ebab07d2f40a@2ndquadrant.com
This adds a flag "deterministic" to collations. If that is false,
such a collation disables various optimizations that assume that
strings are equal only if they are byte-wise equal. That then allows
use cases such as case-insensitive or accent-insensitive comparisons
or handling of strings with different Unicode normal forms.
This functionality is only supported with the ICU provider. At least
glibc doesn't appear to have any locales that work in a
nondeterministic way, so it's not worth supporting this for the libc
provider.
The term "deterministic comparison" in this context is from Unicode
Technical Standard #10
(https://unicode.org/reports/tr10/#Deterministic_Comparison).
This patch makes changes in three areas:
- CREATE COLLATION DDL changes and system catalog changes to support
this new flag.
- Many executor nodes and auxiliary code are extended to track
collations. Previously, this code would just throw away collation
information, because the eventually-called user-defined functions
didn't use it since they only cared about equality, which didn't
need collation information.
- String data type functions that do equality comparisons and hashing
are changed to take the (non-)deterministic flag into account. For
comparison, this just means skipping various shortcuts and tie
breakers that use byte-wise comparison. For hashing, we first need
to convert the input string to a canonical "sort key" using the ICU
analogue of strxfrm().
Reviewed-by: Daniel Verite <daniel@manitou-mail.org>
Reviewed-by: Peter Geoghegan <pg@bowt.ie>
Discussion: https://www.postgresql.org/message-id/flat/1ccc668f-4cbc-0bef-af67-450b47cdfee7@2ndquadrant.com
Too allow table accesses to be not directly dependent on heap, several
new abstractions are needed. Specifically:
1) Heap scans need to be generalized into table scans. Do this by
introducing TableScanDesc, which will be the "base class" for
individual AMs. This contains the AM independent fields from
HeapScanDesc.
The previous heap_{beginscan,rescan,endscan} et al. have been
replaced with a table_ version.
There's no direct replacement for heap_getnext(), as that returned
a HeapTuple, which is undesirable for a other AMs. Instead there's
table_scan_getnextslot(). But note that heap_getnext() lives on,
it's still used widely to access catalog tables.
This is achieved by new scan_begin, scan_end, scan_rescan,
scan_getnextslot callbacks.
2) The portion of parallel scans that's shared between backends need
to be able to do so without the user doing per-AM work. To achieve
that new parallelscan_{estimate, initialize, reinitialize}
callbacks are introduced, which operate on a new
ParallelTableScanDesc, which again can be subclassed by AMs.
As it is likely that several AMs are going to be block oriented,
block oriented callbacks that can be shared between such AMs are
provided and used by heap. table_block_parallelscan_{estimate,
intiialize, reinitialize} as callbacks, and
table_block_parallelscan_{nextpage, init} for use in AMs. These
operate on a ParallelBlockTableScanDesc.
3) Index scans need to be able to access tables to return a tuple, and
there needs to be state across individual accesses to the heap to
store state like buffers. That's now handled by introducing a
sort-of-scan IndexFetchTable, which again is intended to be
subclassed by individual AMs (for heap IndexFetchHeap).
The relevant callbacks for an AM are index_fetch_{end, begin,
reset} to create the necessary state, and index_fetch_tuple to
retrieve an indexed tuple. Note that index_fetch_tuple
implementations need to be smarter than just blindly fetching the
tuples for AMs that have optimizations similar to heap's HOT - the
currently alive tuple in the update chain needs to be fetched if
appropriate.
Similar to table_scan_getnextslot(), it's undesirable to continue
to return HeapTuples. Thus index_fetch_heap (might want to rename
that later) now accepts a slot as an argument. Core code doesn't
have a lot of call sites performing index scans without going
through the systable_* API (in contrast to loads of heap_getnext
calls and working directly with HeapTuples).
Index scans now store the result of a search in
IndexScanDesc->xs_heaptid, rather than xs_ctup->t_self. As the
target is not generally a HeapTuple anymore that seems cleaner.
To be able to sensible adapt code to use the above, two further
callbacks have been introduced:
a) slot_callbacks returns a TupleTableSlotOps* suitable for creating
slots capable of holding a tuple of the AMs
type. table_slot_callbacks() and table_slot_create() are based
upon that, but have additional logic to deal with views, foreign
tables, etc.
While this change could have been done separately, nearly all the
call sites that needed to be adapted for the rest of this commit
also would have been needed to be adapted for
table_slot_callbacks(), making separation not worthwhile.
b) tuple_satisfies_snapshot checks whether the tuple in a slot is
currently visible according to a snapshot. That's required as a few
places now don't have a buffer + HeapTuple around, but a
slot (which in heap's case internally has that information).
Additionally a few infrastructure changes were needed:
I) SysScanDesc, as used by systable_{beginscan, getnext} et al. now
internally uses a slot to keep track of tuples. While
systable_getnext() still returns HeapTuples, and will so for the
foreseeable future, the index API (see 1) above) now only deals with
slots.
The remainder, and largest part, of this commit is then adjusting all
scans in postgres to use the new APIs.
Author: Andres Freund, Haribabu Kommi, Alvaro Herrera
Discussion:
https://postgr.es/m/20180703070645.wchpu5muyto5n647@alap3.anarazel.dehttps://postgr.es/m/20160812231527.GA690404@alvherre.pgsql
We still require AccessExclusiveLock on the partition itself, because
otherwise an insert that violates the newly-imposed partition
constraint could be in progress at the same time that we're changing
that constraint; only the lock level on the parent relation is
weakened.
To make this safe, we have to cope with (at least) three separate
problems. First, relevant DDL might commit while we're in the process
of building a PartitionDesc. If so, find_inheritance_children() might
see a new partition while the RELOID system cache still has the old
partition bound cached, and even before invalidation messages have
been queued. To fix that, if we see that the pg_class tuple seems to
be missing or to have a null relpartbound, refetch the value directly
from the table. We can't get the wrong value, because DETACH PARTITION
still requires AccessExclusiveLock throughout; if we ever want to
change that, this will need more thought. In testing, I found it quite
difficult to hit even the null-relpartbound case; the race condition
is extremely tight, but the theoretical risk is there.
Second, successive calls to RelationGetPartitionDesc might not return
the same answer. The query planner will get confused if lookup up the
PartitionDesc for a particular relation does not return a consistent
answer for the entire duration of query planning. Likewise, query
execution will get confused if the same relation seems to have a
different PartitionDesc at different times. Invent a new
PartitionDirectory concept and use it to ensure consistency. This
ensures that a single invocation of either the planner or the executor
sees the same view of the PartitionDesc from beginning to end, but it
does not guarantee that the planner and the executor see the same
view. Since this allows pointers to old PartitionDesc entries to
survive even after a relcache rebuild, also postpone removing the old
PartitionDesc entry until we're certain no one is using it.
For the most part, it seems to be OK for the planner and executor to
have different views of the PartitionDesc, because the executor will
just ignore any concurrently added partitions which were unknown at
plan time; those partitions won't be part of the inheritance
expansion, but invalidation messages will trigger replanning at some
point. Normally, this happens by the time the very next command is
executed, but if the next command acquires no locks and executes a
prepared query, it can manage not to notice until a new transaction is
started. We might want to tighten that up, but it's material for a
separate patch. There would still be a small window where a query
that started just after an ATTACH PARTITION command committed might
fail to notice its results -- but only if the command starts before
the commit has been acknowledged to the user. All in all, the warts
here around serializability seem small enough to be worth accepting
for the considerable advantage of being able to add partitions without
a full table lock.
Although in general the consequences of new partitions showing up
between planning and execution are limited to the query not noticing
the new partitions, run-time partition pruning will get confused in
that case, so that's the third problem that this patch fixes.
Run-time partition pruning assumes that indexes into the PartitionDesc
are stable between planning and execution. So, add code so that if
new partitions are added between plan time and execution time, the
indexes stored in the subplan_map[] and subpart_map[] arrays within
the plan's PartitionedRelPruneInfo get adjusted accordingly. There
does not seem to be a simple way to generalize this scheme to cope
with partitions that are removed, mostly because they could then get
added back again with different bounds, but it works OK for added
partitions.
This code does not try to ensure that every backend participating in
a parallel query sees the same view of the PartitionDesc. That
currently doesn't matter, because we never pass PartitionDesc
indexes between backends. Each backend will ignore the concurrently
added partitions which it notices, and it doesn't matter if different
backends are ignoring different sets of concurrently added partitions.
If in the future that matters, for example because we allow writes in
parallel query and want all participants to do tuple routing to the same
set of partitions, the PartitionDirectory concept could be improved to
share PartitionDescs across backends. There is a draft patch to
serialize and restore PartitionDescs on the thread where this patch
was discussed, which may be a useful place to start.
Patch by me. Thanks to Alvaro Herrera, David Rowley, Simon Riggs,
Amit Langote, and Michael Paquier for discussion, and to Alvaro
Herrera for some review.
Discussion: http://postgr.es/m/CA+Tgmobt2upbSocvvDej3yzokd7AkiT+PvgFH+a9-5VV1oJNSQ@mail.gmail.com
Discussion: http://postgr.es/m/CA+TgmoZE0r9-cyA-aY6f8WFEROaDLLL7Vf81kZ8MtFCkxpeQSw@mail.gmail.com
Discussion: http://postgr.es/m/CA+TgmoY13KQZF-=HNTrt9UYWYx3_oYOQpu9ioNT49jGgiDpUEA@mail.gmail.com
Until now the the slot to store the conflicting tuple, and the result
of the ON CONFLICT SET, where reused between partitions. That
necessitated changing slots descriptor when switching partitions.
Besides the overhead of switching descriptors on a slot (which
requires memory allocations and prevents JITing), that's importantly
also problematic for tableam. There individual partitions might belong
to different tableams, needing different kinds of slots.
In passing also fix ExecOnConflictUpdate to clear the existing slot at
exit. Otherwise that slot could continue to hold a pin till the query
ends, which could be far too long if the input data set is large, and
there's no further conflicts. While previously also problematic, it's
now more important as there will be more such slots when partitioned.
Author: Andres Freund
Reviewed-By: Robert Haas, David Rowley
Discussion: https://postgr.es/m/20180703070645.wchpu5muyto5n647@alap3.anarazel.de
For the upcoming pluggable table access methods it's quite
inconvenient to store tuples as HeapTuples, as that'd require
converting tuples from a their native format into HeapTuples. Instead
use slots to manage epq tuples.
To fit into that scheme, change the foreign data wrapper callback
RefetchForeignRow, to store the tuple in a slot. Insist on using the
caller provided slot, so it conveniently can be stored in the
corresponding EPQ slot. As there is no in core user of
RefetchForeignRow, that change was done blindly, but we plan to test
that soon.
To avoid duplicating that work for row locks, move row locks to just
directly use the EPQ slots - it previously temporarily stored tuples
in LockRowsState.lr_curtuples, but that doesn't seem beneficial, given
we'd possibly end up with a significant number of additional slots.
The behaviour of es_epqTupleSet[rti -1] is now checked by
es_epqTupleSlot[rti -1] != NULL, as that is distinguishable from a
slot containing an empty tuple.
Author: Andres Freund, Haribabu Kommi, Ashutosh Bapat
Discussion: https://postgr.es/m/20180703070645.wchpu5muyto5n647@alap3.anarazel.de
In preparation for abstracting table storage, convert trigger.c to
track tuples in slots. Which also happens to make code calling
triggers simpler.
As the calling interface for triggers themselves is not changed in
this patch, HeapTuples still are extracted from the slot at that
time. But that's handled solely inside trigger.c, not visible to
callers. It's quite likely that we'll want to revise the external
trigger interface, but that's a separate large project.
As part of this work the slots used for old/new/return tuples are
moved from EState into ResultRelInfo, as different updated tables
might need different slots. The slots are now also now created
on-demand, which is good both from an efficiency POV, but also makes
the modifying code simpler.
Author: Andres Freund, Amit Khandekar and Ashutosh Bapat
Discussion: https://postgr.es/m/20180703070645.wchpu5muyto5n647@alap3.anarazel.de
Before this change FunctionCallInfoData, the struct arguments etc for
V1 function calls are stored in, always had space for
FUNC_MAX_ARGS/100 arguments, storing datums and their nullness in two
arrays. For nearly every function call 100 arguments is far more than
needed, therefore wasting memory. Arg and argnull being two separate
arrays also guarantees that to access a single argument, two
cachelines have to be touched.
Change the layout so there's a single variable-length array with pairs
of value / isnull. That drastically reduces memory consumption for
most function calls (on x86-64 a two argument function now uses
64bytes, previously 936 bytes), and makes it very likely that argument
value and its nullness are on the same cacheline.
Arguments are stored in a new NullableDatum struct, which, due to
padding, needs more memory per argument than before. But as usually
far fewer arguments are stored, and individual arguments are cheaper
to access, that's still a clear win. It's likely that there's other
places where conversion to NullableDatum arrays would make sense,
e.g. TupleTableSlots, but that's for another commit.
Because the function call information is now variable-length
allocations have to take the number of arguments into account. For
heap allocations that can be done with SizeForFunctionCallInfoData(),
for on-stack allocations there's a new LOCAL_FCINFO(name, nargs) macro
that helps to allocate an appropriately sized and aligned variable.
Some places with stack allocation function call information don't know
the number of arguments at compile time, and currently variably sized
stack allocations aren't allowed in postgres. Therefore allow for
FUNC_MAX_ARGS space in these cases. They're not that common, so for
now that seems acceptable.
Because of the need to allocate FunctionCallInfo of the appropriate
size, older extensions may need to update their code. To avoid subtle
breakages, the FunctionCallInfoData struct has been renamed to
FunctionCallInfoBaseData. Most code only references FunctionCallInfo,
so that shouldn't cause much collateral damage.
This change is also a prerequisite for more efficient expression JIT
compilation (by allocating the function call information on the stack,
allowing LLVM to optimize it away); previously the size of the call
information caused problems inside LLVM's optimizer.
Author: Andres Freund
Reviewed-By: Tom Lane
Discussion: https://postgr.es/m/20180605172952.x34m5uz6ju6enaem@alap3.anarazel.de
This is the genam.h equivalent of 4c850ecec6 (which removed
heapam.h from a lot of other headers). There's still a few header
includes of genam.h, but not from central headers anymore.
As a few headers are not indirectly included anymore, execnodes.h and
relscan.h need a few additional includes. Some of the depended on
types were replacable by using the underlying structs, but e.g. for
Snapshot in execnodes.h that'd have gotten more invasive than
reasonable in this commit.
Like the aforementioned commit 4c850ecec6, this requires adding new
genam.h includes to a number of backend files, which likely is also
required in a few external projects.
Author: Andres Freund
Discussion: https://postgr.es/m/20190114000701.y4ttcb74jpskkcfb@alap3.anarazel.de
heapam.h previously was included in a number of widely used
headers (e.g. execnodes.h, indirectly in executor.h, ...). That's
problematic on its own, as heapam.h contains a lot of low-level
details that don't need to be exposed that widely, but becomes more
problematic with the upcoming introduction of pluggable table storage
- it seems inappropriate for heapam.h to be included that widely
afterwards.
heapam.h was largely only included in other headers to get the
HeapScanDesc typedef (which was defined in heapam.h, even though
HeapScanDescData is defined in relscan.h). The better solution here
seems to be to just use the underlying struct (forward declared where
necessary). Similar for BulkInsertState.
Another problem was that LockTupleMode was used in executor.h - parts
of the file tried to cope without heapam.h, but due to the fact that
it indirectly included it, several subsequent violations of that goal
were not not noticed. We could just reuse the approach of declaring
parameters as int, but it seems nicer to move LockTupleMode to
lockoptions.h - that's not a perfect location, but also doesn't seem
bad.
As a number of files relied on implicitly included heapam.h, a
significant number of files grew an explicit include. It's quite
probably that a few external projects will need to do the same.
Author: Andres Freund
Reviewed-By: Alvaro Herrera
Discussion: https://postgr.es/m/20190114000701.y4ttcb74jpskkcfb@alap3.anarazel.de
Previously tables declared WITH OIDS, including a significant fraction
of the catalog tables, stored the oid column not as a normal column,
but as part of the tuple header.
This special column was not shown by default, which was somewhat odd,
as it's often (consider e.g. pg_class.oid) one of the more important
parts of a row. Neither pg_dump nor COPY included the contents of the
oid column by default.
The fact that the oid column was not an ordinary column necessitated a
significant amount of special case code to support oid columns. That
already was painful for the existing, but upcoming work aiming to make
table storage pluggable, would have required expanding and duplicating
that "specialness" significantly.
WITH OIDS has been deprecated since 2005 (commit ff02d0a05280e0).
Remove it.
Removing includes:
- CREATE TABLE and ALTER TABLE syntax for declaring the table to be
WITH OIDS has been removed (WITH (oids[ = true]) will error out)
- pg_dump does not support dumping tables declared WITH OIDS and will
issue a warning when dumping one (and ignore the oid column).
- restoring an pg_dump archive with pg_restore will warn when
restoring a table with oid contents (and ignore the oid column)
- COPY will refuse to load binary dump that includes oids.
- pg_upgrade will error out when encountering tables declared WITH
OIDS, they have to be altered to remove the oid column first.
- Functionality to access the oid of the last inserted row (like
plpgsql's RESULT_OID, spi's SPI_lastoid, ...) has been removed.
The syntax for declaring a table WITHOUT OIDS (or WITH (oids = false)
for CREATE TABLE) is still supported. While that requires a bit of
support code, it seems unnecessary to break applications / dumps that
do not use oids, and are explicit about not using them.
The biggest user of WITH OID columns was postgres' catalog. This
commit changes all 'magic' oid columns to be columns that are normally
declared and stored. To reduce unnecessary query breakage all the
newly added columns are still named 'oid', even if a table's column
naming scheme would indicate 'reloid' or such. This obviously
requires adapting a lot code, mostly replacing oid access via
HeapTupleGetOid() with access to the underlying Form_pg_*->oid column.
The bootstrap process now assigns oids for all oid columns in
genbki.pl that do not have an explicit value (starting at the largest
oid previously used), only oids assigned later by oids will be above
FirstBootstrapObjectId. As the oid column now is a normal column the
special bootstrap syntax for oids has been removed.
Oids are not automatically assigned during insertion anymore, all
backend code explicitly assigns oids with GetNewOidWithIndex(). For
the rare case that insertions into the catalog via SQL are called for
the new pg_nextoid() function can be used (which only works on catalog
tables).
The fact that oid columns on system tables are now normal columns
means that they will be included in the set of columns expanded
by * (i.e. SELECT * FROM pg_class will now include the table's oid,
previously it did not). It'd not technically be hard to hide oid
column by default, but that'd mean confusing behavior would either
have to be carried forward forever, or it'd cause breakage down the
line.
While it's not unlikely that further adjustments are needed, the
scope/invasiveness of the patch makes it worthwhile to get merge this
now. It's painful to maintain externally, too complicated to commit
after the code code freeze, and a dependency of a number of other
patches.
Catversion bump, for obvious reasons.
Author: Andres Freund, with contributions by John Naylor
Discussion: https://postgr.es/m/20180930034810.ywp2c7awz7opzcfr@alap3.anarazel.de
This speeds up write operations (INSERT, UPDATE, DELETE, COPY, as well
as the future MERGE) on partitioned tables.
This changes the setup for tuple routing so that it does far less work
during the initial setup and pushes more work out to when partitions
receive tuples. PartitionDispatchData structs for sub-partitioned
tables are only created when a tuple gets routed through it. The
possibly large arrays in the PartitionTupleRouting struct have largely
been removed. The partitions[] array remains but now never contains any
NULL gaps. Previously the NULLs had to be skipped during
ExecCleanupTupleRouting(), which could add a large overhead to the
cleanup when the number of partitions was large. The partitions[] array
is allocated small to start with and only enlarged when we route tuples
to enough partitions that it runs out of space. This allows us to keep
simple single-row partition INSERTs running quickly. Redesign
The arrays in PartitionTupleRouting which stored the tuple translation maps
have now been removed. These have been moved out into a
PartitionRoutingInfo struct which is an additional field in ResultRelInfo.
The find_all_inheritors() call still remains by far the slowest part of
ExecSetupPartitionTupleRouting(). This commit just removes the other slow
parts.
In passing also rename the tuple translation maps from being ParentToChild
and ChildToParent to being RootToPartition and PartitionToRoot. The old
names mislead you into thinking that a partition of some sub-partitioned
table would translate to the rowtype of the sub-partitioned table rather
than the root partitioned table.
Authors: David Rowley and Amit Langote, heavily revised by Álvaro Herrera
Testing help from Jesper Pedersen and Kato Sho.
Discussion: https://postgr.es/m/CAKJS1f_1RJyFquuCKRFHTdcXqoPX-PYqAd7nz=GVBwvGh4a6xA@mail.gmail.com
Upcoming work intends to allow pluggable ways to introduce new ways of
storing table data. Accessing those table access methods from the
executor requires TupleTableSlots to be carry tuples in the native
format of such storage methods; otherwise there'll be a significant
conversion overhead.
Different access methods will require different data to store tuples
efficiently (just like virtual, minimal, heap already require fields
in TupleTableSlot). To allow that without requiring additional pointer
indirections, we want to have different structs (embedding
TupleTableSlot) for different types of slots. Thus different types of
slots are needed, which requires adapting creators of slots.
The slot that most efficiently can represent a type of tuple in an
executor node will often depend on the type of slot a child node
uses. Therefore we need to track the type of slot is returned by
nodes, so parent slots can create slots based on that.
Relatedly, JIT compilation of tuple deforming needs to know which type
of slot a certain expression refers to, so it can create an
appropriate deforming function for the type of tuple in the slot.
But not all nodes will only return one type of slot, e.g. an append
node will potentially return different types of slots for each of its
subplans.
Therefore add function that allows to query the type of a node's
result slot, and whether it'll always be the same type (whether it's
fixed). This can be queried using ExecGetResultSlotOps().
The scan, result, inner, outer type of slots are automatically
inferred from ExecInitScanTupleSlot(), ExecInitResultSlot(),
left/right subtrees respectively. If that's not correct for a node,
that can be overwritten using new fields in PlanState.
This commit does not introduce the actually abstracted implementation
of different kind of TupleTableSlots, that will be left for a followup
commit. The different types of slots introduced will, for now, still
use the same backing implementation.
While this already partially invalidates the big comment in
tuptable.h, it seems to make more sense to update it later, when the
different TupleTableSlot implementations actually exist.
Author: Ashutosh Bapat and Andres Freund, with changes by Amit Khandekar
Discussion: https://postgr.es/m/20181105210039.hh4vvi4vwoq5ba2q@alap3.anarazel.de
In a lot of nodes the return slot is not required. That can either be
because the node doesn't do any projection (say an Append node), or
because the node does perform projections but the projection is
optimized away because the projection would yield an identical row.
Slots aren't that small, especially for wide rows, so it's worthwhile
to avoid creating them. It's not possible to just skip creating the
slot - it's currently used to determine the tuple descriptor returned
by ExecGetResultType(). So separate the determination of the result
type from the slot creation. The work previously done internally
ExecInitResultTupleSlotTL() can now also be done separately with
ExecInitResultTypeTL() and ExecInitResultSlot(). That way nodes that
aren't guaranteed to need a result slot, can use
ExecInitResultTypeTL() to determine the result type of the node, and
ExecAssignScanProjectionInfo() (via
ExecConditionalAssignProjectionInfo()) determines that a result slot
is needed, it is created with ExecInitResultSlot().
Besides the advantage of avoiding to create slots that then are
unused, this is necessary preparation for later patches around tuple
table slot abstraction. In particular separating the return descriptor
and slot is a prerequisite to allow JITing of tuple deforming with
knowledge of the underlying tuple format, and to avoid unnecessarily
creating JITed tuple deforming for virtual slots.
This commit removes a redundant argument from
ExecInitResultTupleSlotTL(). While this commit touches a lot of the
relevant lines anyway, it'd normally still not worthwhile to cause
breakage, except that aforementioned later commits will touch *all*
ExecInitResultTupleSlotTL() callers anyway (but fits worse
thematically).
Author: Andres Freund
Discussion: https://postgr.es/m/20181105210039.hh4vvi4vwoq5ba2q@alap3.anarazel.de
es_leaf_result_relations doesn't exist; perhaps this was an old name
for es_tuple_routing_result_relations, or maybe this comment has gone
unmaintained through multiple rounds of whacking the code around.
Related comment in execnodes.h was both obsolete and ungrammatical.
If there are many ExecRowMark structs, we spent O(N^2) time in
ExecFindRowMark during executor startup. Once upon a time this was
not of great concern, but the addition of native partitioning has
squeezed out enough other costs that this can become the dominant
overhead in some use-cases for tables with many partitions.
To fix, simply replace that List data structure with an array.
This adds a little bit of cost to execCurrentOf(), but not much,
and anyway that code path is neither of large importance nor very
efficient now. If we ever decide it is a bottleneck, constructing a
hash table for lookup-by-tableoid would likely be the thing to do.
Per complaint from Amit Langote, though this is different from
his fix proposal.
Discussion: https://postgr.es/m/468c85d9-540e-66a2-1dde-fec2b741e688@lab.ntt.co.jp
David Rowley
Jeff Davis
Andres Freund
Tom Lane
Alvaro Herrera
Tomas Vondra
Alexander Korotkov
Peter Eisentraut
Bruce Momjian
Amit Kapila
Michael Paquier
Heikki Linnakangas
Robert Haas