Find file History
petere Fix outdated comment
The issue the comment is referring to was fixed by
Latest commit 3bed67b Jan 19, 2019
Type Name Latest commit message Commit time
Failed to load latest commit information.
Makefile Revert MERGE patch Apr 12, 2018
README Update executor documentation for run-time partition pruning Nov 15, 2018
execAmi.c Update copyright for 2019 Jan 2, 2019
execCurrent.c Don't include genam.h from execnodes.h and relscan.h anymore. Jan 15, 2019
execExpr.c Update copyright for 2019 Jan 2, 2019
execExprInterp.c Update copyright for 2019 Jan 2, 2019
execGrouping.c Update copyright for 2019 Jan 2, 2019
execIndexing.c Don't include genam.h from execnodes.h and relscan.h anymore. Jan 15, 2019
execJunk.c Update copyright for 2019 Jan 2, 2019
execMain.c Don't include heapam.h from others headers. Jan 15, 2019
execParallel.c Update copyright for 2019 Jan 2, 2019
execPartition.c Don't include heapam.h from others headers. Jan 15, 2019
execProcnode.c Update copyright for 2019 Jan 2, 2019
execReplication.c Fix outdated comment Jan 19, 2019
execSRF.c Update copyright for 2019 Jan 2, 2019
execScan.c Update copyright for 2019 Jan 2, 2019
execTuples.c Update copyright for 2019 Jan 2, 2019
execUtils.c Don't include heapam.h from others headers. Jan 15, 2019
functions.c Update copyright for 2019 Jan 2, 2019
instrument.c Update copyright for 2019 Jan 2, 2019
nodeAgg.c Update copyright for 2019 Jan 2, 2019
nodeAppend.c Update copyright for 2019 Jan 2, 2019
nodeBitmapAnd.c Update copyright for 2019 Jan 2, 2019
nodeBitmapHeapscan.c Don't include heapam.h from others headers. Jan 15, 2019
nodeBitmapIndexscan.c Don't include genam.h from execnodes.h and relscan.h anymore. Jan 15, 2019
nodeBitmapOr.c Update copyright for 2019 Jan 2, 2019
nodeCtescan.c Update copyright for 2019 Jan 2, 2019
nodeCustom.c Update copyright for 2019 Jan 2, 2019
nodeForeignscan.c Update copyright for 2019 Jan 2, 2019
nodeFunctionscan.c Update copyright for 2019 Jan 2, 2019
nodeGather.c Update copyright for 2019 Jan 2, 2019
nodeGatherMerge.c Update copyright for 2019 Jan 2, 2019
nodeGroup.c Update copyright for 2019 Jan 2, 2019
nodeHash.c Update copyright for 2019 Jan 2, 2019
nodeHashjoin.c Update copyright for 2019 Jan 2, 2019
nodeIndexonlyscan.c Don't include genam.h from execnodes.h and relscan.h anymore. Jan 15, 2019
nodeIndexscan.c Update copyright for 2019 Jan 2, 2019
nodeLimit.c Update copyright for 2019 Jan 2, 2019
nodeLockRows.c Don't include heapam.h from others headers. Jan 15, 2019
nodeMaterial.c Update copyright for 2019 Jan 2, 2019
nodeMergeAppend.c Update copyright for 2019 Jan 2, 2019
nodeMergejoin.c Update copyright for 2019 Jan 2, 2019
nodeModifyTable.c Fix parent of WCO qual. Jan 15, 2019
nodeNamedtuplestorescan.c Update copyright for 2019 Jan 2, 2019
nodeNestloop.c Update copyright for 2019 Jan 2, 2019
nodeProjectSet.c Update copyright for 2019 Jan 2, 2019
nodeRecursiveunion.c Update copyright for 2019 Jan 2, 2019
nodeResult.c Update copyright for 2019 Jan 2, 2019
nodeSamplescan.c Don't include heapam.h from others headers. Jan 15, 2019
nodeSeqscan.c Don't include heapam.h from others headers. Jan 15, 2019
nodeSetOp.c Update copyright for 2019 Jan 2, 2019
nodeSort.c Update copyright for 2019 Jan 2, 2019
nodeSubplan.c Update copyright for 2019 Jan 2, 2019
nodeSubqueryscan.c Update copyright for 2019 Jan 2, 2019
nodeTableFuncscan.c Update copyright for 2019 Jan 2, 2019
nodeTidscan.c Don't include heapam.h from others headers. Jan 15, 2019
nodeUnique.c Update copyright for 2019 Jan 2, 2019
nodeValuesscan.c Update copyright for 2019 Jan 2, 2019
nodeWindowAgg.c Update copyright for 2019 Jan 2, 2019
nodeWorktablescan.c Update copyright for 2019 Jan 2, 2019
spi.c Update copyright for 2019 Jan 2, 2019
tqueue.c Update copyright for 2019 Jan 2, 2019
tstoreReceiver.c Update copyright for 2019 Jan 2, 2019



The Postgres Executor

The executor processes a tree of "plan nodes".  The plan tree is essentially
a demand-pull pipeline of tuple processing operations.  Each node, when
called, will produce the next tuple in its output sequence, or NULL if no
more tuples are available.  If the node is not a primitive relation-scanning
node, it will have child node(s) that it calls in turn to obtain input

Refinements on this basic model include:

* Choice of scan direction (forwards or backwards).  Caution: this is not
currently well-supported.  It works for primitive scan nodes, but not very
well for joins, aggregates, etc.

* Rescan command to reset a node and make it generate its output sequence
over again.

* Parameters that can alter a node's results.  After adjusting a parameter,
the rescan command must be applied to that node and all nodes above it.
There is a moderately intelligent scheme to avoid rescanning nodes
unnecessarily (for example, Sort does not rescan its input if no parameters
of the input have changed, since it can just reread its stored sorted data).

For a SELECT, it is only necessary to deliver the top-level result tuples
to the client.  For INSERT/UPDATE/DELETE, the actual table modification
operations happen in a top-level ModifyTable plan node.  If the query
includes a RETURNING clause, the ModifyTable node delivers the computed
RETURNING rows as output, otherwise it returns nothing.  Handling INSERT
is pretty straightforward: the tuples returned from the plan tree below
ModifyTable are inserted into the correct result relation.  For UPDATE,
the plan tree returns the computed tuples to be updated, plus a "junk"
(hidden) CTID column identifying which table row is to be replaced by each
one.  For DELETE, the plan tree need only deliver a CTID column, and the
ModifyTable node visits each of those rows and marks the row deleted.

XXX a great deal more documentation needs to be written here...

Plan Trees and State Trees

The plan tree delivered by the planner contains a tree of Plan nodes (struct
types derived from struct Plan).  During executor startup we build a parallel
tree of identical structure containing executor state nodes --- generally,
every plan node type has a corresponding executor state node type.  Each node
in the state tree has a pointer to its corresponding node in the plan tree,
plus executor state data as needed to implement that node type.  This
arrangement allows the plan tree to be completely read-only so far as the
executor is concerned: all data that is modified during execution is in the
state tree.  Read-only plan trees make life much simpler for plan caching and

A corresponding executor state node may not be created during executor startup
if the executor determines that an entire subplan is not required due to
execution time partition pruning determining that no matching records will be
found there.  This currently only occurs for Append and MergeAppend nodes.  In
this case the non-required subplans are ignored and the executor state's
subnode array will become out of sequence to the plan's subplan list.

Each Plan node may have expression trees associated with it, to represent
its target list, qualification conditions, etc.  These trees are also
read-only to the executor, but the executor state for expression evaluation
does not mirror the Plan expression's tree shape, as explained below.
Rather, there's just one ExprState node per expression tree, although this
may have sub-nodes for some complex expression node types.

Altogether there are four classes of nodes used in these trees: Plan nodes,
their corresponding PlanState nodes, Expr nodes, and ExprState nodes.
(Actually, there are also List nodes, which are used as "glue" in all
three tree-based representations.)

Expression Trees and ExprState nodes

Expression trees, in contrast to Plan trees, are not mirrored into a
corresponding tree of state nodes.  Instead each separately executable
expression tree (e.g. a Plan's qual or targetlist) is represented by one
ExprState node.  The ExprState node contains the information needed to
evaluate the expression in a compact, linear form.  That compact form is
stored as a flat array in ExprState->steps[] (an array of ExprEvalStep,
not ExprEvalStep *).

The reasons for choosing such a representation include:
- commonly the amount of work needed to evaluate one Expr-type node is
  small enough that the overhead of having to perform a tree-walk
  during evaluation is significant.
- the flat representation can be evaluated non-recursively within a single
  function, reducing stack depth and function call overhead.
- such a representation is usable both for fast interpreted execution,
  and for compiling into native code.

The Plan-tree representation of an expression is compiled into an
ExprState node by ExecInitExpr().  As much complexity as possible should
be handled by ExecInitExpr() (and helpers), instead of execution time
where both interpreted and compiled versions would need to deal with the
complexity.  Besides duplicating effort between execution approaches,
runtime initialization checks also have a small but noticeable cost every
time the expression is evaluated.  Therefore, we allow ExecInitExpr() to
precompute information that we do not expect to vary across execution of a
single query, for example the set of CHECK constraint expressions to be
applied to a domain type.  This could not be done at plan time without
greatly increasing the number of events that require plan invalidation.
(Previously, some information of this kind was rechecked on each
expression evaluation, but that seems like unnecessary overhead.)

Expression Initialization

During ExecInitExpr() and similar routines, Expr trees are converted
into the flat representation.  Each Expr node might be represented by
zero, one, or more ExprEvalSteps.

Each ExprEvalStep's work is determined by its opcode (of enum ExprEvalOp)
and it stores the result of its work into the Datum variable and boolean
null flag variable pointed to by ExprEvalStep->resvalue/resnull.
Complex expressions are performed by chaining together several steps.
For example, "a + b" (one OpExpr, with two Var expressions) would be
represented as two steps to fetch the Var values, and one step for the
evaluation of the function underlying the + operator.  The steps for the
Vars would have their resvalue/resnull pointing directly to the appropriate
arg[] and argnull[] array elements in the FunctionCallInfoData struct that
is used by the function evaluation step, thus avoiding extra work to copy
the result values around.

The last entry in a completed ExprState->steps array is always an
EEOP_DONE step; this removes the need to test for end-of-array while
iterating.  Also, if the expression contains any variable references (to
user columns of the ExprContext's INNER, OUTER, or SCAN tuples), the steps
array begins with EEOP_*_FETCHSOME steps that ensure that the relevant
tuples have been deconstructed to make the required columns directly
available (cf. slot_getsomeattrs()).  This allows individual Var-fetching
steps to be little more than an array lookup.

Most of ExecInitExpr()'s work is done by the recursive function
ExecInitExprRec() and its subroutines.  ExecInitExprRec() maps one Expr
node into the steps required for execution, recursing as needed for

Each ExecInitExprRec() call has to specify where that subexpression's
results are to be stored (via the resv/resnull parameters).  This allows
the above scenario of evaluating a (sub-)expression directly into
fcinfo->arg/argnull, but also requires some care: target Datum/isnull
variables may not be shared with another ExecInitExprRec() unless the
results are only needed by steps executing before further usages of those
target Datum/isnull variables.  Due to the non-recursiveness of the
ExprEvalStep representation that's usually easy to guarantee.

ExecInitExprRec() pushes new operations into the ExprState->steps array
using ExprEvalPushStep().  To keep the steps as a consecutively laid out
array, ExprEvalPushStep() has to repalloc the entire array when there's
not enough space.  Because of that it is *not* allowed to point directly
into any of the steps during expression initialization.  Therefore, the
resv/resnull for a subexpression usually point to some storage that is
palloc'd separately from the steps array.  For instance, the
FunctionCallInfoData for a function call step is separately allocated
rather than being part of the ExprEvalStep array.  The overall result
of a complete expression is typically returned into the resvalue/resnull
fields of the ExprState node itself.

Some steps, e.g. boolean expressions, allow skipping evaluation of
certain subexpressions.  In the flat representation this amounts to
jumping to some later step rather than just continuing consecutively
with the next step.  The target for such a jump is represented by
the integer index in the ExprState->steps array of the step to execute
next.  (Compare the EEO_NEXT and EEO_JUMP macros in execExprInterp.c.)

Typically, ExecInitExprRec() has to push a jumping step into the steps
array, then recursively generate steps for the subexpression that might
get skipped over, then go back and fix up the jump target index using
the now-known length of the subexpression's steps.  This is handled by
adjust_jumps lists in execExpr.c.

The last step in constructing an ExprState is to apply ExecReadyExpr(),
which readies it for execution using whichever execution method has been

Expression Evaluation

To allow for different methods of expression evaluation, and for
better branch/jump target prediction, expressions are evaluated by
calling ExprState->evalfunc (via ExprEvalExpr() and friends).

ExprReadyExpr() can choose the method of interpretation by setting
evalfunc to an appropriate function.  The default execution function,
ExecInterpExpr, is implemented in execExprInterp.c; see its header
comment for details.  Special-case evalfuncs are used for certain
especially-simple expressions.

Note that a lot of the more complex expression evaluation steps, which are
less performance-critical than the simpler ones, are implemented as
separate functions outside the fast-path of expression execution, allowing
their implementation to be shared between interpreted and compiled
expression evaluation.  This means that these helper functions are not
allowed to perform expression step dispatch themselves, as the method of
dispatch will vary based on the caller.  The helpers therefore cannot call
for the execution of subexpressions; all subexpression results they need
must be computed by earlier steps.  And dispatch to the following
expression step must be performed after returning from the helper.

Targetlist Evaluation

ExecBuildProjectionInfo builds an ExprState that has the effect of
evaluating a targetlist into ExprState->resultslot.  A generic targetlist
expression is executed by evaluating it as discussed above (storing the
result into the ExprState's resvalue/resnull fields) and then using an
EEOP_ASSIGN_TMP step to move the result into the appropriate tts_values[]
and tts_isnull[] array elements of the result slot.  There are special
fast-path step types (EEOP_ASSIGN_*_VAR) to handle targetlist entries that
are simple Vars using only one step instead of two.

Memory Management

A "per query" memory context is created during CreateExecutorState();
all storage allocated during an executor invocation is allocated in that
context or a child context.  This allows easy reclamation of storage
during executor shutdown --- rather than messing with retail pfree's and
probable storage leaks, we just destroy the memory context.

In particular, the plan state trees and expression state trees described
in the previous section are allocated in the per-query memory context.

To avoid intra-query memory leaks, most processing while a query runs
is done in "per tuple" memory contexts, which are so-called because they
are typically reset to empty once per tuple.  Per-tuple contexts are usually
associated with ExprContexts, and commonly each PlanState node has its own
ExprContext to evaluate its qual and targetlist expressions in.

Query Processing Control Flow

This is a sketch of control flow for full query processing:


			creates per-query context
		switch to per-query context to run ExecInitNode
		ExecInitNode --- recursively scans plan tree
				recurse into subsidiary nodes
				creates per-tuple context

		ExecProcNode --- recursively called in per-query context
			ExecEvalExpr --- called in per-tuple context
			ResetExprContext --- to free memory

		ExecPostprocessPlan --- run any unfinished ModifyTable nodes

		ExecEndNode --- recursively releases resources
			frees per-query context and child contexts


Per above comments, it's not really critical for ExecEndNode to free any
memory; it'll all go away in FreeExecutorState anyway.  However, we do need to
be careful to close relations, drop buffer pins, etc, so we do need to scan
the plan state tree to find these sorts of resources.

The executor can also be used to evaluate simple expressions without any Plan
tree ("simple" meaning "no aggregates and no sub-selects", though such might
be hidden inside function calls).  This case has a flow of control like

		creates per-query context

	CreateExprContext	-- or use GetPerTupleExprContext(estate)
		creates per-tuple context

		temporarily switch to per-query context
		run the expression through expression_planner

	Repeatedly do:
			ExecEvalExpr --- called in per-tuple context
		ResetExprContext --- to free memory

		frees per-query context, as well as ExprContext
		(a separate FreeExprContext call is not necessary)

EvalPlanQual (READ COMMITTED Update Checking)

For simple SELECTs, the executor need only pay attention to tuples that are
valid according to the snapshot seen by the current transaction (ie, they
were inserted by a previously committed transaction, and not deleted by any
previously committed transaction).  However, for UPDATE and DELETE it is not
cool to modify or delete a tuple that's been modified by an open or
concurrently-committed transaction.  If we are running in SERIALIZABLE
isolation level then we just raise an error when this condition is seen to
occur.  In READ COMMITTED isolation level, we must work a lot harder.

The basic idea in READ COMMITTED mode is to take the modified tuple
committed by the concurrent transaction (after waiting for it to commit,
if need be) and re-evaluate the query qualifications to see if it would
still meet the quals.  If so, we regenerate the updated tuple (if we are
doing an UPDATE) from the modified tuple, and finally update/delete the
modified tuple.  SELECT FOR UPDATE/SHARE behaves similarly, except that its
action is just to lock the modified tuple and return results based on that
version of the tuple.

To implement this checking, we actually re-run the query from scratch for
each modified tuple (or set of tuples, for SELECT FOR UPDATE), with the
relation scan nodes tweaked to return only the current tuples --- either
the original ones, or the updated (and now locked) versions of the modified
tuple(s).  If this query returns a tuple, then the modified tuple(s) pass
the quals (and the query output is the suitably modified update tuple, if
we're doing UPDATE).  If no tuple is returned, then the modified tuple(s)
fail the quals, so we ignore the current result tuple and continue the
original query.

In UPDATE/DELETE, only the target relation needs to be handled this way.
In SELECT FOR UPDATE, there may be multiple relations flagged FOR UPDATE,
so we obtain lock on the current tuple version in each such relation before
executing the recheck.

It is also possible that there are relations in the query that are not
to be locked (they are neither the UPDATE/DELETE target nor specified to
be locked in SELECT FOR UPDATE/SHARE).  When re-running the test query
we want to use the same rows from these relations that were joined to
the locked rows.  For ordinary relations this can be implemented relatively
cheaply by including the row TID in the join outputs and re-fetching that
TID.  (The re-fetch is expensive, but we're trying to optimize the normal
case where no re-test is needed.)  We have also to consider non-table
relations, such as a ValuesScan or FunctionScan.  For these, since there
is no equivalent of TID, the only practical solution seems to be to include
the entire row value in the join output row.

We disallow set-returning functions in the targetlist of SELECT FOR UPDATE,
so as to ensure that at most one tuple can be returned for any particular
set of scan tuples.  Otherwise we'd get duplicates due to the original
query returning the same set of scan tuples multiple times.  Likewise,
SRFs are disallowed in an UPDATE's targetlist.  There, they would have the
effect of the same row being updated multiple times, which is not very
useful --- and updates after the first would have no effect anyway.