/* ----------------------------------------------------------------
* ExecInitBitmapHeapScan
*
* Initializes the scan's state information.
* ----------------------------------------------------------------
*/
BitmapHeapScanState *
ExecInitBitmapHeapScan(BitmapHeapScan *node, EState *estate, int eflags)
{
BitmapHeapScanState *scanstate;
Relation currentRelation;
int io_concurrency;
/* check for unsupported flags */
Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
/*
* Assert caller didn't ask for an unsafe snapshot --- see comments at
* head of file.
*/
Assert(IsMVCCSnapshot(estate->es_snapshot));
/*
* create state structure
*/
scanstate = makeNode(BitmapHeapScanState);
scanstate->ss.ps.plan = (Plan *) node;
scanstate->ss.ps.state = estate;
scanstate->ss.ps.ExecProcNode = ExecBitmapHeapScan;
scanstate->tbm = NULL;
scanstate->tbmiterator = NULL;
scanstate->tbmres = NULL;
scanstate->exact_pages = 0;
scanstate->lossy_pages = 0;
scanstate->prefetch_iterator = NULL;
scanstate->prefetch_pages = 0;
scanstate->prefetch_target = 0;
/* may be updated below */
scanstate->prefetch_maximum = target_prefetch_pages;
scanstate->pscan_len = 0;
scanstate->initialized = false;
scanstate->shared_tbmiterator = NULL;
scanstate->pstate = NULL;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &scanstate->ss.ps);
/*
* initialize child expressions
*/
scanstate->ss.ps.qual =
ExecInitQual(node->scan.plan.qual, (PlanState *) scanstate);
scanstate->bitmapqualorig =
ExecInitQual(node->bitmapqualorig, (PlanState *) scanstate);
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &scanstate->ss.ps);
ExecInitScanTupleSlot(estate, &scanstate->ss);
/*
* open the base relation and acquire appropriate lock on it.
*/
currentRelation = ExecOpenScanRelation(estate, node->scan.scanrelid, eflags);
/*
* Determine the maximum for prefetch_target. If the tablespace has a
* specific IO concurrency set, use that to compute the corresponding
* maximum value; otherwise, we already initialized to the value computed
* by the GUC machinery.
*/
io_concurrency =
get_tablespace_io_concurrency(currentRelation->rd_rel->reltablespace);
if (io_concurrency != effective_io_concurrency)
{
double maximum;
if (ComputeIoConcurrency(io_concurrency, &maximum))
scanstate->prefetch_maximum = rint(maximum);
}
scanstate->ss.ss_currentRelation = currentRelation;
/*
* Even though we aren't going to do a conventional seqscan, it is useful
* to create a HeapScanDesc --- most of the fields in it are usable.
*/
scanstate->ss.ss_currentScanDesc = heap_beginscan_bm(currentRelation,
estate->es_snapshot,
0,
NULL);
/*
* get the scan type from the relation descriptor.
*/
ExecAssignScanType(&scanstate->ss, RelationGetDescr(currentRelation));
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&scanstate->ss.ps);
ExecAssignScanProjectionInfo(&scanstate->ss);
/*
* initialize child nodes
*
* We do this last because the child nodes will open indexscans on our
* relation's indexes, and we want to be sure we have acquired a lock on
* the relation first.
*/
outerPlanState(scanstate) = ExecInitNode(outerPlan(node), estate, eflags);
/*
* all done.
*/
return scanstate;
}
/* ----------------------------------------------------------------
* ExecInitFunctionScan
* ----------------------------------------------------------------
*/
FunctionScanState *
ExecInitFunctionScan(FunctionScan *node, EState *estate, int eflags)
{
FunctionScanState *scanstate;
RangeTblEntry *rte;
Oid funcrettype;
TypeFuncClass functypclass;
TupleDesc tupdesc = NULL;
/*
* FunctionScan should not have any children.
*/
Assert(outerPlan(node) == NULL);
Assert(innerPlan(node) == NULL);
/*
* create new ScanState for node
*/
scanstate = makeNode(FunctionScanState);
scanstate->ss.ps.plan = (Plan *) node;
scanstate->ss.ps.state = estate;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &scanstate->ss.ps);
#define FUNCTIONSCAN_NSLOTS 2
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &scanstate->ss.ps);
ExecInitScanTupleSlot(estate, &scanstate->ss);
/*
* initialize child expressions
*/
scanstate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->scan.plan.targetlist,
(PlanState *) scanstate);
scanstate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->scan.plan.qual,
(PlanState *) scanstate);
/*
* get info about function
*/
rte = rt_fetch(node->scan.scanrelid, estate->es_range_table);
Assert(rte->rtekind == RTE_FUNCTION);
/*
* Now determine if the function returns a simple or composite type, and
* build an appropriate tupdesc.
*/
functypclass = get_expr_result_type(rte->funcexpr,
&funcrettype,
&tupdesc);
if (functypclass == TYPEFUNC_COMPOSITE)
{
/* Composite data type, e.g. a table's row type */
Assert(tupdesc);
/* Must copy it out of typcache for safety */
tupdesc = CreateTupleDescCopy(tupdesc);
}
else if (functypclass == TYPEFUNC_SCALAR)
{
/* Base data type, i.e. scalar */
char *attname = strVal(linitial(rte->eref->colnames));
tupdesc = CreateTemplateTupleDesc(1, false);
TupleDescInitEntry(tupdesc,
(AttrNumber) 1,
attname,
funcrettype,
-1,
0);
}
else if (functypclass == TYPEFUNC_RECORD)
{
tupdesc = BuildDescForRelation(rte->coldeflist);
}
else
{
/* crummy error message, but parser should have caught this */
elog(ERROR, "function in FROM has unsupported return type");
}
/*
* For RECORD results, make sure a typmod has been assigned. (The
* function should do this for itself, but let's cover things in case it
* doesn't.)
*/
BlessTupleDesc(tupdesc);
scanstate->tupdesc = tupdesc;
ExecAssignScanType(&scanstate->ss, tupdesc, false);
/*
* Other node-specific setup
*/
scanstate->tuplestorestate = NULL;
scanstate->funcexpr = ExecInitExpr((Expr *) rte->funcexpr,
(PlanState *) scanstate);
scanstate->ss.ps.ps_TupFromTlist = false;
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&scanstate->ss.ps);
ExecAssignScanProjectionInfo(&scanstate->ss);
return scanstate;
}
/* ----------------------------------------------------------------
* ExecHashTableCreate
*
* create an empty hashtable data structure for hashjoin.
* ----------------------------------------------------------------
*/
HashJoinTable
ExecHashTableCreate(Hash *node, List *hashOperators, bool keepNulls)
{
HashJoinTable hashtable;
Plan *outerNode;
int nbuckets;
int nbatch;
int num_skew_mcvs;
int log2_nbuckets;
int nkeys;
int i;
ListCell *ho;
MemoryContext oldcxt;
/*
* Get information about the size of the relation to be hashed (it's the
* "outer" subtree of this node, but the inner relation of the hashjoin).
* Compute the appropriate size of the hash table.
*/
outerNode = outerPlan(node);
ExecChooseHashTableSize(outerNode->plan_rows, outerNode->plan_width,
OidIsValid(node->skewTable),
&nbuckets, &nbatch, &num_skew_mcvs);
/* nbuckets must be a power of 2 */
log2_nbuckets = my_log2(nbuckets);
Assert(nbuckets == (1 << log2_nbuckets));
/*
* Initialize the hash table control block.
*
* The hashtable control block is just palloc'd from the executor's
* per-query memory context. Everything else should be kept inside the
* subsidiary hashCxt or batchCxt.
*/
hashtable = (HashJoinTable) palloc(sizeof(HashJoinTableData));
hashtable->nbuckets = nbuckets;
hashtable->nbuckets_original = nbuckets;
hashtable->nbuckets_optimal = nbuckets;
hashtable->log2_nbuckets = log2_nbuckets;
hashtable->log2_nbuckets_optimal = log2_nbuckets;
hashtable->buckets = NULL;
hashtable->keepNulls = keepNulls;
hashtable->skewEnabled = false;
hashtable->skewBucket = NULL;
hashtable->skewBucketLen = 0;
hashtable->nSkewBuckets = 0;
hashtable->skewBucketNums = NULL;
hashtable->nbatch = nbatch;
hashtable->curbatch = 0;
hashtable->nbatch_original = nbatch;
hashtable->nbatch_outstart = nbatch;
hashtable->growEnabled = true;
hashtable->totalTuples = 0;
hashtable->skewTuples = 0;
hashtable->innerBatchFile = NULL;
hashtable->outerBatchFile = NULL;
hashtable->spaceUsed = 0;
hashtable->spacePeak = 0;
hashtable->spaceAllowed = work_mem * 1024L;
hashtable->spaceUsedSkew = 0;
hashtable->spaceAllowedSkew =
hashtable->spaceAllowed * SKEW_WORK_MEM_PERCENT / 100;
hashtable->chunks = NULL;
#ifdef HJDEBUG
printf("Hashjoin %p: initial nbatch = %d, nbuckets = %d\n",
hashtable, nbatch, nbuckets);
#endif
/*
* Create temporary memory contexts in which to keep the hashtable working
* storage. See notes in executor/hashjoin.h.
*/
hashtable->hashCxt = AllocSetContextCreate(CurrentMemoryContext,
"HashTableContext",
ALLOCSET_DEFAULT_SIZES);
hashtable->batchCxt = AllocSetContextCreate(hashtable->hashCxt,
"HashBatchContext",
ALLOCSET_DEFAULT_SIZES);
/* Allocate data that will live for the life of the hashjoin */
oldcxt = MemoryContextSwitchTo(hashtable->hashCxt);
/*
* Get info about the hash functions to be used for each hash key. Also
* remember whether the join operators are strict.
*/
nkeys = list_length(hashOperators);
hashtable->outer_hashfunctions =
(FmgrInfo *) palloc(nkeys * sizeof(FmgrInfo));
hashtable->inner_hashfunctions =
(FmgrInfo *) palloc(nkeys * sizeof(FmgrInfo));
hashtable->hashStrict = (bool *) palloc(nkeys * sizeof(bool));
i = 0;
foreach(ho, hashOperators)
{
Oid hashop = lfirst_oid(ho);
Oid left_hashfn;
Oid right_hashfn;
if (!get_op_hash_functions(hashop, &left_hashfn, &right_hashfn))
elog(ERROR, "could not find hash function for hash operator %u",
hashop);
fmgr_info(left_hashfn, &hashtable->outer_hashfunctions[i]);
fmgr_info(right_hashfn, &hashtable->inner_hashfunctions[i]);
hashtable->hashStrict[i] = op_strict(hashop);
i++;
}
/* ----------------------------------------------------------------
* ExecInitExternalScan
* ----------------------------------------------------------------
*/
ExternalScanState *
ExecInitExternalScan(ExternalScan *node, EState *estate, int eflags)
{
ResultRelSegFileInfo *segfileinfo = NULL;
ExternalScanState *externalstate;
Relation currentRelation;
FileScanDesc currentScanDesc;
Assert(outerPlan(node) == NULL);
Assert(innerPlan(node) == NULL);
/*
* create state structure
*/
externalstate = makeNode(ExternalScanState);
externalstate->ss.ps.plan = (Plan *) node;
externalstate->ss.ps.state = estate;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &externalstate->ss.ps);
/*
* initialize child expressions
*/
externalstate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->scan.plan.targetlist,
(PlanState *) externalstate);
externalstate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->scan.plan.qual,
(PlanState *) externalstate);
/* Check if targetlist or qual contains a var node referencing the ctid column */
externalstate->cdb_want_ctid = contain_ctid_var_reference(&node->scan);
ItemPointerSetInvalid(&externalstate->cdb_fake_ctid);
#define EXTSCAN_NSLOTS 2
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &externalstate->ss.ps);
ExecInitScanTupleSlot(estate, &externalstate->ss);
/*
* get the relation object id from the relid'th entry in the range table
* and open that relation.
*/
currentRelation = ExecOpenScanExternalRelation(estate, node->scan.scanrelid);
if (Gp_role == GP_ROLE_EXECUTE && node->err_aosegfileinfos)
{
segfileinfo = (ResultRelSegFileInfo *)list_nth(node->err_aosegfileinfos, GetQEIndex());
}
else
{
segfileinfo = NULL;
}
currentScanDesc = external_beginscan(currentRelation,
node->scan.scanrelid,
node->scancounter,
node->uriList,
node->fmtOpts,
node->fmtType,
node->isMasterOnly,
node->rejLimit,
node->rejLimitInRows,
node->fmterrtbl,
segfileinfo,
node->encoding,
node->scan.plan.qual);
externalstate->ss.ss_currentRelation = currentRelation;
externalstate->ess_ScanDesc = currentScanDesc;
ExecAssignScanType(&externalstate->ss, RelationGetDescr(currentRelation));
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&externalstate->ss.ps);
ExecAssignScanProjectionInfo(&externalstate->ss);
/*
* If eflag contains EXEC_FLAG_REWIND or EXEC_FLAG_BACKWARD or EXEC_FLAG_MARK,
* then this node is not eager free safe.
*/
externalstate->ss.ps.delayEagerFree =
((eflags & (EXEC_FLAG_REWIND | EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)) != 0);
initGpmonPktForExternalScan((Plan *)node, &externalstate->ss.ps.gpmon_pkt, estate);
return externalstate;
}
/* Return number of TupleTableSlots used by nodeDML.*/
int
ExecCountSlotsDML(DML *node)
{
return ExecCountSlotsNode(outerPlan(node)) + DML_NSLOTS;
}
/* ----------------------------------------------------------------
* ExecInitSetOp
*
* This initializes the setop node state structures and
* the node's subplan.
* ----------------------------------------------------------------
*/
SetOpState *
ExecInitSetOp(SetOp *node, EState *estate, int eflags)
{
SetOpState *setopstate;
/* check for unsupported flags */
Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
/*
* create state structure
*/
setopstate = makeNode(SetOpState);
setopstate->ps.plan = (Plan *) node;
setopstate->ps.state = estate;
setopstate->eqfunctions = NULL;
setopstate->hashfunctions = NULL;
setopstate->setop_done = false;
setopstate->numOutput = 0;
setopstate->pergroup = NULL;
setopstate->grp_firstTuple = NULL;
setopstate->hashtable = NULL;
setopstate->tableContext = NULL;
/*
* Miscellaneous initialization
*
* SetOp nodes have no ExprContext initialization because they never call
* ExecQual or ExecProject. But they do need a per-tuple memory context
* anyway for calling execTuplesMatch.
*/
setopstate->tempContext =
AllocSetContextCreate(CurrentMemoryContext,
"SetOp",
ALLOCSET_DEFAULT_SIZES);
/*
* If hashing, we also need a longer-lived context to store the hash
* table. The table can't just be kept in the per-query context because
* we want to be able to throw it away in ExecReScanSetOp.
*/
if (node->strategy == SETOP_HASHED)
setopstate->tableContext =
AllocSetContextCreate(CurrentMemoryContext,
"SetOp hash table",
ALLOCSET_DEFAULT_SIZES);
/*
* Tuple table initialization
*/
ExecInitResultTupleSlot(estate, &setopstate->ps);
/*
* initialize child nodes
*
* If we are hashing then the child plan does not need to handle REWIND
* efficiently; see ExecReScanSetOp.
*/
if (node->strategy == SETOP_HASHED)
eflags &= ~EXEC_FLAG_REWIND;
outerPlanState(setopstate) = ExecInitNode(outerPlan(node), estate, eflags);
/*
* setop nodes do no projections, so initialize projection info for this
* node appropriately
*/
ExecAssignResultTypeFromTL(&setopstate->ps);
setopstate->ps.ps_ProjInfo = NULL;
/*
* Precompute fmgr lookup data for inner loop. We need both equality and
* hashing functions to do it by hashing, but only equality if not
* hashing.
*/
if (node->strategy == SETOP_HASHED)
execTuplesHashPrepare(node->numCols,
node->dupOperators,
&setopstate->eqfunctions,
&setopstate->hashfunctions);
else
setopstate->eqfunctions =
execTuplesMatchPrepare(node->numCols,
node->dupOperators);
if (node->strategy == SETOP_HASHED)
{
build_hash_table(setopstate);
setopstate->table_filled = false;
}
else
{
setopstate->pergroup =
(SetOpStatePerGroup) palloc0(sizeof(SetOpStatePerGroupData));
}
return setopstate;
}
/* ----------------------------------------------------------------
* ExecInitCteScan
* ----------------------------------------------------------------
*/
CteScanState *
ExecInitCteScan(CteScan *node, EState *estate, int eflags)
{
CteScanState *scanstate;
ParamExecData *prmdata;
/* check for unsupported flags */
Assert(!(eflags & EXEC_FLAG_MARK));
/*
* For the moment we have to force the tuplestore to allow REWIND, because
* we might be asked to rescan the CTE even though upper levels didn't
* tell us to be prepared to do it efficiently. Annoying, since this
* prevents truncation of the tuplestore. XXX FIXME
*/
eflags |= EXEC_FLAG_REWIND;
/*
* CteScan should not have any children.
*/
Assert(outerPlan(node) == NULL);
Assert(innerPlan(node) == NULL);
/*
* create new CteScanState for node
*/
scanstate = makeNode(CteScanState);
scanstate->ss.ps.plan = (Plan *) node;
scanstate->ss.ps.state = estate;
scanstate->eflags = eflags;
scanstate->cte_table = NULL;
scanstate->eof_cte = false;
/*
* Find the already-initialized plan for the CTE query.
*/
scanstate->cteplanstate = (PlanState *) list_nth(estate->es_subplanstates,
node->ctePlanId - 1);
/*
* The Param slot associated with the CTE query is used to hold a pointer
* to the CteState of the first CteScan node that initializes for this
* CTE. This node will be the one that holds the shared state for all the
* CTEs, particularly the shared tuplestore.
*/
prmdata = &(estate->es_param_exec_vals[node->cteParam]);
Assert(prmdata->execPlan == NULL);
Assert(!prmdata->isnull);
scanstate->leader = (CteScanState *) DatumGetPointer(prmdata->value);
if (scanstate->leader == NULL)
{
/* I am the leader */
prmdata->value = PointerGetDatum(scanstate);
scanstate->leader = scanstate;
scanstate->cte_table = tuplestore_begin_heap(true, false, work_mem);
tuplestore_set_eflags(scanstate->cte_table, scanstate->eflags);
scanstate->readptr = 0;
}
else
{
/* Not the leader */
Assert(IsA(scanstate->leader, CteScanState));
scanstate->readptr =
tuplestore_alloc_read_pointer(scanstate->leader->cte_table,
scanstate->eflags);
}
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &scanstate->ss.ps);
/*
* initialize child expressions
*/
scanstate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->scan.plan.targetlist,
(PlanState *) scanstate);
scanstate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->scan.plan.qual,
(PlanState *) scanstate);
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &scanstate->ss.ps);
ExecInitScanTupleSlot(estate, &scanstate->ss);
/*
* The scan tuple type (ie, the rowtype we expect to find in the work
* table) is the same as the result rowtype of the CTE query.
*/
ExecAssignScanType(&scanstate->ss,
ExecGetResultType(scanstate->cteplanstate));
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&scanstate->ss.ps);
ExecAssignScanProjectionInfo(&scanstate->ss);
scanstate->ss.ps.ps_TupFromTlist = false;
return scanstate;
}
/* ----------------------------------------------------------------
* ExecInitSubqueryScan
* ----------------------------------------------------------------
*/
SubqueryScanState *
ExecInitSubqueryScan(SubqueryScan *node, EState *estate, int eflags)
{
SubqueryScanState *subquerystate;
/* check for unsupported flags */
Assert(!(eflags & EXEC_FLAG_MARK));
/*
* SubqueryScan should not have any "normal" children. Also, if planner
* left anything in subrtable, it's fishy.
*/
Assert(outerPlan(node) == NULL);
Assert(innerPlan(node) == NULL);
Assert(node->subrtable == NIL);
/*
* Since subquery nodes create its own executor state,
* and pass it down to its child nodes, we always
* initialize the subquery node. However, some
* fields are not initialized if not necessary, see
* below.
*/
/*
* create state structure
*/
subquerystate = makeNode(SubqueryScanState);
subquerystate->ss.ps.plan = (Plan *) node;
subquerystate->ss.ps.state = estate;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &subquerystate->ss.ps);
/*
* initialize child expressions
*/
subquerystate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->scan.plan.targetlist,
(PlanState *) subquerystate);
subquerystate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->scan.plan.qual,
(PlanState *) subquerystate);
/* Check if targetlist or qual contains a var node referencing the ctid column */
subquerystate->cdb_want_ctid = contain_ctid_var_reference(&node->scan);
ItemPointerSetInvalid(&subquerystate->cdb_fake_ctid);
#define SUBQUERYSCAN_NSLOTS 2
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &subquerystate->ss.ps);
ExecInitScanTupleSlot(estate, &subquerystate->ss);
/*
* initialize subquery
*/
subquerystate->subplan = ExecInitNode(node->subplan, estate, eflags);
/* return borrowed share node list */
estate->es_sharenode = estate->es_sharenode;
/*subquerystate->ss.ps.ps_TupFromTlist = false;*/
/*
* Initialize scan tuple type (needed by ExecAssignScanProjectionInfo)
*/
ExecAssignScanType(&subquerystate->ss,
ExecGetResultType(subquerystate->subplan));
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&subquerystate->ss.ps);
ExecAssignScanProjectionInfo(&subquerystate->ss);
initGpmonPktForSubqueryScan((Plan *)node, &subquerystate->ss.ps.gpmon_pkt, estate);
return subquerystate;
}
/* ----------------------------------------------------------------
* ExecInitForeignScan
* ----------------------------------------------------------------
*/
ForeignScanState *
ExecInitForeignScan(ForeignScan *node, EState *estate, int eflags)
{
ForeignScanState *scanstate;
Relation currentRelation = NULL;
Index scanrelid = node->scan.scanrelid;
Index tlistvarno;
FdwRoutine *fdwroutine;
/* check for unsupported flags */
Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
/*
* create state structure
*/
scanstate = makeNode(ForeignScanState);
scanstate->ss.ps.plan = (Plan *) node;
scanstate->ss.ps.state = estate;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &scanstate->ss.ps);
scanstate->ss.ps.ps_TupFromTlist = false;
/*
* initialize child expressions
*/
scanstate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->scan.plan.targetlist,
(PlanState *) scanstate);
scanstate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->scan.plan.qual,
(PlanState *) scanstate);
scanstate->fdw_recheck_quals = (List *)
ExecInitExpr((Expr *) node->fdw_recheck_quals,
(PlanState *) scanstate);
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &scanstate->ss.ps);
ExecInitScanTupleSlot(estate, &scanstate->ss);
/*
* open the base relation, if any, and acquire an appropriate lock on it;
* also acquire function pointers from the FDW's handler
*/
if (scanrelid > 0)
{
currentRelation = ExecOpenScanRelation(estate, scanrelid, eflags);
scanstate->ss.ss_currentRelation = currentRelation;
fdwroutine = GetFdwRoutineForRelation(currentRelation, true);
}
else
{
/* We can't use the relcache, so get fdwroutine the hard way */
fdwroutine = GetFdwRoutineByServerId(node->fs_server);
}
/*
* Determine the scan tuple type. If the FDW provided a targetlist
* describing the scan tuples, use that; else use base relation's rowtype.
*/
if (node->fdw_scan_tlist != NIL || currentRelation == NULL)
{
TupleDesc scan_tupdesc;
scan_tupdesc = ExecTypeFromTL(node->fdw_scan_tlist, false);
ExecAssignScanType(&scanstate->ss, scan_tupdesc);
/* Node's targetlist will contain Vars with varno = INDEX_VAR */
tlistvarno = INDEX_VAR;
}
else
{
ExecAssignScanType(&scanstate->ss, RelationGetDescr(currentRelation));
/* Node's targetlist will contain Vars with varno = scanrelid */
tlistvarno = scanrelid;
}
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&scanstate->ss.ps);
ExecAssignScanProjectionInfoWithVarno(&scanstate->ss, tlistvarno);
/*
* Initialize FDW-related state.
*/
scanstate->fdwroutine = fdwroutine;
scanstate->fdw_state = NULL;
/* Initialize any outer plan. */
if (outerPlan(node))
outerPlanState(scanstate) =
ExecInitNode(outerPlan(node), estate, eflags);
/*
* Tell the FDW to initialize the scan.
*/
fdwroutine->BeginForeignScan(scanstate, eflags);
return scanstate;
}
/* ----------------------------------------------------------------
* ExecInitRecursiveUnionScan
* ----------------------------------------------------------------
*/
RecursiveUnionState *
ExecInitRecursiveUnion(RecursiveUnion *node, EState *estate, int eflags)
{
RecursiveUnionState *rustate;
ParamExecData *prmdata;
/* check for unsupported flags */
Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
/*
* create state structure
*/
rustate = makeNode(RecursiveUnionState);
rustate->ps.plan = (Plan *) node;
rustate->ps.state = estate;
rustate->eqfunctions = NULL;
rustate->hashfunctions = NULL;
rustate->hashtable = NULL;
rustate->tempContext = NULL;
rustate->tableContext = NULL;
/* initialize processing state */
rustate->recursing = false;
rustate->intermediate_empty = true;
rustate->working_table = tuplestore_begin_heap(false, false, work_mem);
rustate->intermediate_table = tuplestore_begin_heap(false, false, work_mem);
/*
* If hashing, we need a per-tuple memory context for comparisons, and a
* longer-lived context to store the hash table. The table can't just be
* kept in the per-query context because we want to be able to throw it
* away when rescanning.
*/
if (node->numCols > 0)
{
rustate->tempContext =
AllocSetContextCreate(CurrentMemoryContext,
"RecursiveUnion",
ALLOCSET_DEFAULT_SIZES);
rustate->tableContext =
AllocSetContextCreate(CurrentMemoryContext,
"RecursiveUnion hash table",
ALLOCSET_DEFAULT_SIZES);
}
/*
* Make the state structure available to descendant WorkTableScan nodes
* via the Param slot reserved for it.
*/
prmdata = &(estate->es_param_exec_vals[node->wtParam]);
Assert(prmdata->execPlan == NULL);
prmdata->value = PointerGetDatum(rustate);
prmdata->isnull = false;
/*
* Miscellaneous initialization
*
* RecursiveUnion plans don't have expression contexts because they never
* call ExecQual or ExecProject.
*/
Assert(node->plan.qual == NIL);
/*
* RecursiveUnion nodes still have Result slots, which hold pointers to
* tuples, so we have to initialize them.
*/
ExecInitResultTupleSlot(estate, &rustate->ps);
/*
* Initialize result tuple type and projection info. (Note: we have to
* set up the result type before initializing child nodes, because
* nodeWorktablescan.c expects it to be valid.)
*/
ExecAssignResultTypeFromTL(&rustate->ps);
rustate->ps.ps_ProjInfo = NULL;
/*
* initialize child nodes
*/
outerPlanState(rustate) = ExecInitNode(outerPlan(node), estate, eflags);
innerPlanState(rustate) = ExecInitNode(innerPlan(node), estate, eflags);
/*
* If hashing, precompute fmgr lookup data for inner loop, and create the
* hash table.
*/
if (node->numCols > 0)
{
execTuplesHashPrepare(node->numCols,
node->dupOperators,
&rustate->eqfunctions,
&rustate->hashfunctions);
build_hash_table(rustate);
}
return rustate;
}
/* ----------------------------------------------------------------
* ExecInitSubqueryScan
* ----------------------------------------------------------------
*/
SubqueryScanState *
ExecInitSubqueryScan(SubqueryScan *node, EState *estate, int eflags)
{
SubqueryScanState *subquerystate;
/* check for unsupported flags */
Assert(!(eflags & EXEC_FLAG_MARK));
/* SubqueryScan should not have any "normal" children */
Assert(outerPlan(node) == NULL);
Assert(innerPlan(node) == NULL);
/*
* create state structure
*/
subquerystate = makeNode(SubqueryScanState);
subquerystate->ss.ps.plan = (Plan *) node;
subquerystate->ss.ps.state = estate;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &subquerystate->ss.ps);
/*
* initialize child expressions
*/
subquerystate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->scan.plan.targetlist,
(PlanState *) subquerystate);
subquerystate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->scan.plan.qual,
(PlanState *) subquerystate);
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &subquerystate->ss.ps);
ExecInitScanTupleSlot(estate, &subquerystate->ss);
/*
* initialize subquery
*/
subquerystate->subplan = ExecInitNode(node->subplan, estate, eflags);
subquerystate->ss.ps.ps_TupFromTlist = false;
/*
* Initialize scan tuple type (needed by ExecAssignScanProjectionInfo)
*/
ExecAssignScanType(&subquerystate->ss,
ExecGetResultType(subquerystate->subplan));
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&subquerystate->ss.ps);
ExecAssignScanProjectionInfo(&subquerystate->ss);
return subquerystate;
}
/* ----------------------------------------------------------------
* ExecInitResult
*
* Creates the run-time state information for the result node
* produced by the planner and initializes outer relations
* (child nodes).
* ----------------------------------------------------------------
*/
ResultState *
ExecInitResult(Result *node, EState *estate, int eflags)
{
ResultState *resstate;
/* check for unsupported flags */
Assert(!(eflags & (EXEC_FLAG_MARK | EXEC_FLAG_BACKWARD)) ||
outerPlan(node) != NULL);
/*
* create state structure
*/
resstate = makeNode(ResultState);
resstate->ps.plan = (Plan *) node;
resstate->ps.state = estate;
resstate->rs_done = false;
resstate->rs_checkqual = (node->resconstantqual == NULL) ? false : true;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &resstate->ps);
resstate->ps.ps_TupFromTlist = false;
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &resstate->ps);
/*
* initialize child expressions
*/
resstate->ps.targetlist = (List *)
ExecInitExpr((Expr *) node->plan.targetlist,
(PlanState *) resstate);
resstate->ps.qual = (List *)
ExecInitExpr((Expr *) node->plan.qual,
(PlanState *) resstate);
resstate->resconstantqual = ExecInitExpr((Expr *) node->resconstantqual,
(PlanState *) resstate);
/*
* initialize child nodes
*/
outerPlanState(resstate) = ExecInitNode(outerPlan(node), estate, eflags);
/*
* we don't use inner plan
*/
Assert(innerPlan(node) == NULL);
/*
* initialize tuple type and projection info
*/
ExecAssignResultTypeFromTL(&resstate->ps);
ExecAssignProjectionInfo(&resstate->ps, NULL);
return resstate;
}
/* ----------------------------------------------------------------
* ExecHash
*
* build hash table for hashjoin, all do partitioning if more
* than one batches are required.
* ----------------------------------------------------------------
*/
TupleTableSlot *
ExecHash(Hash *node)
{
EState *estate;
HashState *hashstate;
Plan *outerNode;
Var *hashkey;
HashJoinTable hashtable;
TupleTableSlot *slot;
ExprContext *econtext;
int nbatch;
File *batches;
RelativeAddr *batchPos;
int *batchSizes;
int i;
RelativeAddr *innerbatchNames;
/* ----------------
* get state info from node
* ----------------
*/
hashstate = node->hashstate;
estate = node->plan.state;
outerNode = outerPlan(node);
hashtable = node->hashtable;
if (hashtable == NULL)
elog(WARN, "ExecHash: hash table is NULL.");
nbatch = hashtable->nbatch;
if (nbatch > 0) { /* if needs hash partition */
innerbatchNames = (RelativeAddr *) ABSADDR(hashtable->innerbatchNames);
/* --------------
* allocate space for the file descriptors of batch files
* then open the batch files in the current processes.
* --------------
*/
batches = (File*)palloc(nbatch * sizeof(File));
for (i=0; i<nbatch; i++) {
batches[i] = FileNameOpenFile(ABSADDR(innerbatchNames[i]),
O_CREAT | O_RDWR, 0600);
}
hashstate->hashBatches = batches;
batchPos = (RelativeAddr*) ABSADDR(hashtable->innerbatchPos);
batchSizes = (int*) ABSADDR(hashtable->innerbatchSizes);
}
/* ----------------
* set expression context
* ----------------
*/
hashkey = node->hashkey;
econtext = hashstate->cstate.cs_ExprContext;
/* ----------------
* get tuple and insert into the hash table
* ----------------
*/
for (;;) {
slot = ExecProcNode(outerNode, (Plan*)node);
if (TupIsNull(slot))
break;
econtext->ecxt_innertuple = slot;
ExecHashTableInsert(hashtable, econtext, hashkey,
hashstate->hashBatches);
ExecClearTuple(slot);
}
/*
* end of build phase, flush all the last pages of the batches.
*/
for (i=0; i<nbatch; i++) {
if (FileSeek(batches[i], 0L, SEEK_END) < 0)
perror("FileSeek");
if (FileWrite(batches[i],ABSADDR(hashtable->batch)+i*BLCKSZ,BLCKSZ) < 0)
perror("FileWrite");
NDirectFileWrite++;
}
/* ---------------------
* Return the slot so that we have the tuple descriptor
* when we need to save/restore them. -Jeff 11 July 1991
* ---------------------
*/
return slot;
}
HashJoinTable
ExecHashTableCreate(Hash *node)
{
Plan *outerNode;
int nbatch;
int ntuples;
int tupsize;
IpcMemoryId shmid;
HashJoinTable hashtable;
HashBucket bucket;
int nbuckets;
int totalbuckets;
int bucketsize;
int i;
RelativeAddr *outerbatchNames;
RelativeAddr *outerbatchPos;
RelativeAddr *innerbatchNames;
RelativeAddr *innerbatchPos;
int *innerbatchSizes;
RelativeAddr tempname;
nbatch = -1;
HashTBSize = NBuffers/2;
while (nbatch < 0) {
/*
* determine number of batches for the hashjoin
*/
HashTBSize *= 2;
nbatch = ExecHashPartition(node);
}
/* ----------------
* get information about the size of the relation
* ----------------
*/
outerNode = outerPlan(node);
ntuples = outerNode->plan_size;
if (ntuples <= 0)
ntuples = 1000; /* XXX just a hack */
tupsize = outerNode->plan_width + sizeof(HeapTupleData);
/*
* totalbuckets is the total number of hash buckets needed for
* the entire relation
*/
totalbuckets = ceil((double)ntuples/NTUP_PER_BUCKET);
bucketsize = LONGALIGN (NTUP_PER_BUCKET * tupsize + sizeof(*bucket));
/*
* nbuckets is the number of hash buckets for the first pass
* of hybrid hashjoin
*/
nbuckets = (HashTBSize - nbatch) * BLCKSZ / (bucketsize * FUDGE_FAC);
if (totalbuckets < nbuckets)
totalbuckets = nbuckets;
if (nbatch == 0)
nbuckets = totalbuckets;
#ifdef HJDEBUG
printf("nbatch = %d, totalbuckets = %d, nbuckets = %d\n", nbatch, totalbuckets, nbuckets);
#endif
/* ----------------
* in non-parallel machines, we don't need to put the hash table
* in the shared memory. We just palloc it.
* ----------------
*/
hashtable = (HashJoinTable)palloc((HashTBSize+1)*BLCKSZ);
shmid = 0;
if (hashtable == NULL) {
elog(WARN, "not enough memory for hashjoin.");
}
/* ----------------
* initialize the hash table header
* ----------------
*/
hashtable->nbuckets = nbuckets;
hashtable->totalbuckets = totalbuckets;
hashtable->bucketsize = bucketsize;
hashtable->shmid = shmid;
hashtable->top = sizeof(HashTableData);
hashtable->bottom = HashTBSize * BLCKSZ;
/*
* hashtable->readbuf has to be long aligned!!!
*/
hashtable->readbuf = hashtable->bottom;
hashtable->nbatch = nbatch;
hashtable->curbatch = 0;
hashtable->pcount = hashtable->nprocess = 0;
if (nbatch > 0) {
/* ---------------
* allocate and initialize the outer batches
* ---------------
*/
outerbatchNames = (RelativeAddr*)ABSADDR(
hashTableAlloc(nbatch * sizeof(RelativeAddr), hashtable));
outerbatchPos = (RelativeAddr*)ABSADDR(
hashTableAlloc(nbatch * sizeof(RelativeAddr), hashtable));
for (i=0; i<nbatch; i++) {
tempname = hashTableAlloc(12, hashtable);
mk_hj_temp(ABSADDR(tempname));
outerbatchNames[i] = tempname;
outerbatchPos[i] = -1;
}
hashtable->outerbatchNames = RELADDR(outerbatchNames);
hashtable->outerbatchPos = RELADDR(outerbatchPos);
/* ---------------
* allocate and initialize the inner batches
* ---------------
*/
innerbatchNames = (RelativeAddr*)ABSADDR(
hashTableAlloc(nbatch * sizeof(RelativeAddr), hashtable));
innerbatchPos = (RelativeAddr*)ABSADDR(
hashTableAlloc(nbatch * sizeof(RelativeAddr), hashtable));
innerbatchSizes = (int*)ABSADDR(
hashTableAlloc(nbatch * sizeof(int), hashtable));
for (i=0; i<nbatch; i++) {
tempname = hashTableAlloc(12, hashtable);
mk_hj_temp(ABSADDR(tempname));
innerbatchNames[i] = tempname;
innerbatchPos[i] = -1;
innerbatchSizes[i] = 0;
}
hashtable->innerbatchNames = RELADDR(innerbatchNames);
hashtable->innerbatchPos = RELADDR(innerbatchPos);
hashtable->innerbatchSizes = RELADDR(innerbatchSizes);
}
else {
hashtable->outerbatchNames = (RelativeAddr)NULL;
hashtable->outerbatchPos = (RelativeAddr)NULL;
hashtable->innerbatchNames = (RelativeAddr)NULL;
hashtable->innerbatchPos = (RelativeAddr)NULL;
hashtable->innerbatchSizes = (RelativeAddr)NULL;
}
hashtable->batch = (RelativeAddr)LONGALIGN(hashtable->top +
bucketsize * nbuckets);
hashtable->overflownext=hashtable->batch + nbatch * BLCKSZ;
/* ----------------
* initialize each hash bucket
* ----------------
*/
bucket = (HashBucket)ABSADDR(hashtable->top);
for (i=0; i<nbuckets; i++) {
bucket->top = RELADDR((char*)bucket + sizeof(*bucket));
bucket->bottom = bucket->top;
bucket->firstotuple = bucket->lastotuple = -1;
bucket = (HashBucket)LONGALIGN(((char*)bucket + bucketsize));
}
return(hashtable);
}
int
ExecCountSlotsGroup(Group *node)
{
return ExecCountSlotsNode(outerPlan(node)) + GROUP_NSLOTS;
}
/* ----------------------------------------------------------------
* ExecInitBitmapHeapScan
*
* Initializes the scan's state information.
* ----------------------------------------------------------------
*/
BitmapHeapScanState *
ExecInitBitmapHeapScan(BitmapHeapScan *node, EState *estate, int eflags)
{
BitmapHeapScanState *scanstate;
Relation currentRelation;
/* check for unsupported flags */
Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
/*
* Assert caller didn't ask for an unsafe snapshot --- see comments at
* head of file.
*/
Assert(IsMVCCSnapshot(estate->es_snapshot));
/*
* create state structure
*/
scanstate = makeNode(BitmapHeapScanState);
scanstate->ss.ps.plan = (Plan *) node;
scanstate->ss.ps.state = estate;
scanstate->tbm = NULL;
scanstate->tbmres = NULL;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &scanstate->ss.ps);
scanstate->ss.ps.ps_TupFromTlist = false;
/*
* initialize child expressions
*/
scanstate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->scan.plan.targetlist,
(PlanState *) scanstate);
scanstate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->scan.plan.qual,
(PlanState *) scanstate);
scanstate->bitmapqualorig = (List *)
ExecInitExpr((Expr *) node->bitmapqualorig,
(PlanState *) scanstate);
#define BITMAPHEAPSCAN_NSLOTS 2
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &scanstate->ss.ps);
ExecInitScanTupleSlot(estate, &scanstate->ss);
/*
* open the base relation and acquire appropriate lock on it.
*/
currentRelation = ExecOpenScanRelation(estate, node->scan.scanrelid);
scanstate->ss.ss_currentRelation = currentRelation;
/*
* Even though we aren't going to do a conventional seqscan, it is useful
* to create a HeapScanDesc --- this checks the relation size and sets up
* statistical infrastructure for us.
*/
scanstate->ss.ss_currentScanDesc = heap_beginscan(currentRelation,
estate->es_snapshot,
0,
NULL);
/*
* One problem is that heap_beginscan counts a "sequential scan" start,
* when we actually aren't doing any such thing. Reverse out the added
* scan count. (Eventually we may want to count bitmap scans separately.)
*/
pgstat_discount_heap_scan(&scanstate->ss.ss_currentScanDesc->rs_pgstat_info);
/*
* get the scan type from the relation descriptor.
*/
ExecAssignScanType(&scanstate->ss, RelationGetDescr(currentRelation));
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&scanstate->ss.ps);
ExecAssignScanProjectionInfo(&scanstate->ss);
/*
* initialize child nodes
*
* We do this last because the child nodes will open indexscans on our
* relation's indexes, and we want to be sure we have acquired a lock on
* the relation first.
*/
outerPlanState(scanstate) = ExecInitNode(outerPlan(node), estate, eflags);
/*
* all done.
*/
return scanstate;
}
/* ----------------------------------------------------------------
* ExecHashTableCreate
*
* create an empty hashtable data structure for hashjoin.
* ----------------------------------------------------------------
*/
HashJoinTable
ExecHashTableCreate(Hash *node, List *hashOperators)
{
HashJoinTable hashtable;
Plan *outerNode;
int nbuckets;
int nbatch;
int nkeys;
int i;
ListCell *ho;
MemoryContext oldcxt;
/*
* Get information about the size of the relation to be hashed (it's the
* "outer" subtree of this node, but the inner relation of the hashjoin).
* Compute the appropriate size of the hash table.
*/
outerNode = outerPlan(node);
ExecChooseHashTableSize(outerNode->plan_rows, outerNode->plan_width,
&nbuckets, &nbatch);
#ifdef HJDEBUG
printf("nbatch = %d, nbuckets = %d\n", nbatch, nbuckets);
#endif
/*
* Initialize the hash table control block.
*
* The hashtable control block is just palloc'd from the executor's
* per-query memory context.
*/
hashtable = (HashJoinTable) palloc(sizeof(HashJoinTableData));
hashtable->nbuckets = nbuckets;
hashtable->buckets = NULL;
hashtable->nbatch = nbatch;
hashtable->curbatch = 0;
hashtable->nbatch_original = nbatch;
hashtable->nbatch_outstart = nbatch;
hashtable->growEnabled = true;
hashtable->totalTuples = 0;
hashtable->innerBatchFile = NULL;
hashtable->outerBatchFile = NULL;
hashtable->spaceUsed = 0;
hashtable->spaceAllowed = work_mem * 1024L;
/*
* Get info about the hash functions to be used for each hash key.
*/
nkeys = list_length(hashOperators);
hashtable->hashfunctions = (FmgrInfo *) palloc(nkeys * sizeof(FmgrInfo));
i = 0;
foreach(ho, hashOperators)
{
Oid hashfn;
hashfn = get_op_hash_function(lfirst_oid(ho));
if (!OidIsValid(hashfn))
elog(ERROR, "could not find hash function for hash operator %u",
lfirst_oid(ho));
fmgr_info(hashfn, &hashtable->hashfunctions[i]);
i++;
}
int
ExecCountSlotsBitmapHeapScan(BitmapHeapScan *node)
{
return ExecCountSlotsNode(outerPlan((Plan *) node)) +
ExecCountSlotsNode(innerPlan((Plan *) node)) + BITMAPHEAPSCAN_NSLOTS;
}
int
ExecCountSlotsResult(Result *node)
{
return ExecCountSlotsNode(outerPlan(node)) + RESULT_NSLOTS;
}
/* -----------------
* ExecInitGroup
*
* Creates the run-time information for the group node produced by the
* planner and initializes its outer subtree
* -----------------
*/
GroupState *
ExecInitGroup(Group *node, EState *estate)
{
GroupState *grpstate;
/*
* create state structure
*/
grpstate = makeNode(GroupState);
grpstate->ss.ps.plan = (Plan *) node;
grpstate->ss.ps.state = estate;
grpstate->grp_firstTuple = NULL;
grpstate->grp_done = FALSE;
/*
* create expression context
*/
ExecAssignExprContext(estate, &grpstate->ss.ps);
#define GROUP_NSLOTS 2
/*
* tuple table initialization
*/
ExecInitScanTupleSlot(estate, &grpstate->ss);
ExecInitResultTupleSlot(estate, &grpstate->ss.ps);
/*
* initialize child expressions
*/
grpstate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->plan.targetlist,
(PlanState *) grpstate);
grpstate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->plan.qual,
(PlanState *) grpstate);
/*
* initialize child nodes
*/
outerPlanState(grpstate) = ExecInitNode(outerPlan(node), estate);
/*
* initialize tuple type.
*/
ExecAssignScanTypeFromOuterPlan(&grpstate->ss);
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&grpstate->ss.ps);
ExecAssignProjectionInfo(&grpstate->ss.ps);
/*
* Precompute fmgr lookup data for inner loop
*/
grpstate->eqfunctions =
execTuplesMatchPrepare(ExecGetScanType(&grpstate->ss),
node->numCols,
node->grpColIdx);
return grpstate;
}
/* ----------------------------------------------------------------
* ExecInitSeqScan
* ----------------------------------------------------------------
*/
SeqScanState *
ExecInitSeqScan(SeqScan *node, EState *estate, int eflags)
{
SeqScanState *scanstate;
/*
* Once upon a time it was possible to have an outerPlan of a SeqScan, but
* not any more.
*/
Assert(outerPlan(node) == NULL);
Assert(innerPlan(node) == NULL);
/*
* create state structure
*/
scanstate = makeNode(SeqScanState);
scanstate->ps.plan = (Plan *) node;
scanstate->ps.state = estate;
scanstate->ss->scan_state = SCAN_INIT;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &scanstate->ps);
/*
* initialize child expressions
*/
scanstate->ps.targetlist = (List *)
ExecInitExpr((Expr *) node->plan.targetlist,
(PlanState *) scanstate);
scanstate->ps.qual = (List *)
ExecInitExpr((Expr *) node->plan.qual,
(PlanState *) scanstate);
#define SEQSCAN_NSLOTS 2
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &scanstate->ps);
ExecInitScanTupleSlot(estate, scanstate);
/*
* initialize scan relation
*/
InitScanRelation(scanstate, estate);
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&scanstate->ps);
ExecAssignScanProjectionInfo(scanstate);
initGpmonPktForSeqScan((Plan *)node, &scanstate->ps.gpmon_pkt, estate);
/*
* If eflag contains EXEC_FLAG_REWIND or EXEC_FLAG_BACKWARD or EXEC_FLAG_MARK,
* then this node is not eager free safe.
*/
scanstate->ps.delayEagerFree =
((eflags & (EXEC_FLAG_REWIND | EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)) != 0);
return scanstate;
}
/* ----------------------------------------------------------------
* ExecInitBitmapHeapScan
*
* Initializes the scan's state information.
* ----------------------------------------------------------------
*/
BitmapHeapScanState *
ExecInitBitmapHeapScan(BitmapHeapScan *node, EState *estate, int eflags)
{
BitmapHeapScanState *scanstate;
Relation currentRelation;
/* check for unsupported flags */
Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
/*
* Assert caller didn't ask for an unsafe snapshot --- see comments at
* head of file.
*/
Assert(IsMVCCSnapshot(estate->es_snapshot));
/*
* create state structure
*/
scanstate = makeNode(BitmapHeapScanState);
scanstate->ss.ps.plan = (Plan *) node;
scanstate->ss.ps.state = estate;
scanstate->tbm = NULL;
scanstate->tbmiterator = NULL;
scanstate->tbmres = NULL;
scanstate->exact_pages = 0;
scanstate->lossy_pages = 0;
scanstate->prefetch_iterator = NULL;
scanstate->prefetch_pages = 0;
scanstate->prefetch_target = 0;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &scanstate->ss.ps);
scanstate->ss.ps.ps_TupFromTlist = false;
/*
* initialize child expressions
*/
scanstate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->scan.plan.targetlist,
(PlanState *) scanstate);
scanstate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->scan.plan.qual,
(PlanState *) scanstate);
scanstate->bitmapqualorig = (List *)
ExecInitExpr((Expr *) node->bitmapqualorig,
(PlanState *) scanstate);
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &scanstate->ss.ps);
ExecInitScanTupleSlot(estate, &scanstate->ss);
/*
* open the base relation and acquire appropriate lock on it.
*/
currentRelation = ExecOpenScanRelation(estate, node->scan.scanrelid, eflags);
scanstate->ss.ss_currentRelation = currentRelation;
/*
* Even though we aren't going to do a conventional seqscan, it is useful
* to create a HeapScanDesc --- most of the fields in it are usable.
*/
scanstate->ss.ss_currentScanDesc = heap_beginscan_bm(currentRelation,
estate->es_snapshot,
0,
NULL);
/*
* get the scan type from the relation descriptor.
*/
ExecAssignScanType(&scanstate->ss, RelationGetDescr(currentRelation));
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&scanstate->ss.ps);
ExecAssignScanProjectionInfo(&scanstate->ss);
/*
* initialize child nodes
*
* We do this last because the child nodes will open indexscans on our
* relation's indexes, and we want to be sure we have acquired a lock on
* the relation first.
*/
outerPlanState(scanstate) = ExecInitNode(outerPlan(node), estate, eflags);
/*
* all done.
*/
return scanstate;
}
/* ----------------------------------------------------------------
* ExecInitNestLoop
* ----------------------------------------------------------------
*/
NestLoopState *
ExecInitNestLoop(NestLoop *node, EState *estate)
{
NestLoopState *nlstate;
NL1_printf("ExecInitNestLoop: %s\n",
"initializing node");
/*
* create state structure
*/
nlstate = makeNode(NestLoopState);
nlstate->js.ps.plan = (Plan *) node;
nlstate->js.ps.state = estate;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &nlstate->js.ps);
/*
* initialize child expressions
*/
nlstate->js.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->join.plan.targetlist,
(PlanState *) nlstate);
nlstate->js.ps.qual = (List *)
ExecInitExpr((Expr *) node->join.plan.qual,
(PlanState *) nlstate);
nlstate->js.jointype = node->join.jointype;
nlstate->js.joinqual = (List *)
ExecInitExpr((Expr *) node->join.joinqual,
(PlanState *) nlstate);
/*
* initialize child nodes
*/
outerPlanState(nlstate) = ExecInitNode(outerPlan(node), estate);
innerPlanState(nlstate) = ExecInitNode(innerPlan(node), estate);
#define NESTLOOP_NSLOTS 2
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &nlstate->js.ps);
switch (node->join.jointype)
{
case JOIN_INNER:
case JOIN_IN:
break;
case JOIN_LEFT:
nlstate->nl_NullInnerTupleSlot =
ExecInitNullTupleSlot(estate,
ExecGetResultType(innerPlanState(nlstate)));
break;
default:
elog(ERROR, "unrecognized join type: %d",
(int) node->join.jointype);
}
/*
* initialize tuple type and projection info
*/
ExecAssignResultTypeFromTL(&nlstate->js.ps);
ExecAssignProjectionInfo(&nlstate->js.ps);
/*
* finally, wipe the current outer tuple clean.
*/
nlstate->js.ps.ps_OuterTupleSlot = NULL;
nlstate->js.ps.ps_TupFromTlist = false;
nlstate->nl_NeedNewOuter = true;
nlstate->nl_MatchedOuter = false;
NL1_printf("ExecInitNestLoop: %s\n",
"node initialized");
return nlstate;
}
/* Returns the number of slots needed for this operator */
int
ExecCountSlotsBitmapTableScan(BitmapTableScan *node)
{
return ExecCountSlotsNode(outerPlan((Plan *) node)) +
ExecCountSlotsNode(innerPlan((Plan *) node)) + BITMAPTABLESCAN_NSLOTS;
}
/* ----------------------------------------------------------------
* ExecInitGather
* ----------------------------------------------------------------
*/
GatherState *
ExecInitGather(Gather *node, EState *estate, int eflags)
{
GatherState *gatherstate;
Plan *outerNode;
bool hasoid;
TupleDesc tupDesc;
/* Gather node doesn't have innerPlan node. */
Assert(innerPlan(node) == NULL);
/*
* create state structure
*/
gatherstate = makeNode(GatherState);
gatherstate->ps.plan = (Plan *) node;
gatherstate->ps.state = estate;
gatherstate->need_to_scan_locally = !node->single_copy;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &gatherstate->ps);
/*
* initialize child expressions
*/
gatherstate->ps.targetlist = (List *)
ExecInitExpr((Expr *) node->plan.targetlist,
(PlanState *) gatherstate);
gatherstate->ps.qual = (List *)
ExecInitExpr((Expr *) node->plan.qual,
(PlanState *) gatherstate);
/*
* tuple table initialization
*/
gatherstate->funnel_slot = ExecInitExtraTupleSlot(estate);
ExecInitResultTupleSlot(estate, &gatherstate->ps);
/*
* now initialize outer plan
*/
outerNode = outerPlan(node);
outerPlanState(gatherstate) = ExecInitNode(outerNode, estate, eflags);
gatherstate->ps.ps_TupFromTlist = false;
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&gatherstate->ps);
ExecAssignProjectionInfo(&gatherstate->ps, NULL);
/*
* Initialize funnel slot to same tuple descriptor as outer plan.
*/
if (!ExecContextForcesOids(&gatherstate->ps, &hasoid))
hasoid = false;
tupDesc = ExecTypeFromTL(outerNode->targetlist, hasoid);
ExecSetSlotDescriptor(gatherstate->funnel_slot, tupDesc);
return gatherstate;
}
int
ExecCountSlotsBitmapIndexScan(BitmapIndexScan *node)
{
return ExecCountSlotsNode(outerPlan((Plan *) node)) +
ExecCountSlotsNode(innerPlan((Plan *) node)) + BITMAPINDEXSCAN_NSLOTS;
}
/* ----------------------------------------------------------------
* ExecHashTableCreate
*
* create an empty hashtable data structure for hashjoin.
* ----------------------------------------------------------------
*/
HashJoinTable
ExecHashTableCreate(HashState *hashState, HashJoinState *hjstate, List *hashOperators, uint64 operatorMemKB)
{
HashJoinTable hashtable;
Plan *outerNode;
int nbuckets;
int nbatch;
int nkeys;
int i;
ListCell *ho;
MemoryContext oldcxt;
START_MEMORY_ACCOUNT(hashState->ps.memoryAccount);
{
Hash *node = (Hash *) hashState->ps.plan;
/*
* Get information about the size of the relation to be hashed (it's the
* "outer" subtree of this node, but the inner relation of the hashjoin).
* Compute the appropriate size of the hash table.
*/
outerNode = outerPlan(node);
/*
* Initialize the hash table control block.
*
* The hashtable control block is just palloc'd from the executor's
* per-query memory context.
*/
hashtable = (HashJoinTable)palloc0(sizeof(HashJoinTableData));
hashtable->buckets = NULL;
hashtable->bloom = NULL;
hashtable->curbatch = 0;
hashtable->growEnabled = true;
hashtable->totalTuples = 0;
hashtable->batches = NULL;
hashtable->work_set = NULL;
hashtable->state_file = NULL;
hashtable->spaceAllowed = operatorMemKB * 1024L;
hashtable->stats = NULL;
hashtable->eagerlyReleased = false;
hashtable->hjstate = hjstate;
/*
* Create temporary memory contexts in which to keep the hashtable working
* storage. See notes in executor/hashjoin.h.
*/
hashtable->hashCxt = AllocSetContextCreate(CurrentMemoryContext,
"HashTableContext",
ALLOCSET_DEFAULT_MINSIZE,
ALLOCSET_DEFAULT_INITSIZE,
ALLOCSET_DEFAULT_MAXSIZE);
hashtable->batchCxt = AllocSetContextCreate(hashtable->hashCxt,
"HashBatchContext",
ALLOCSET_DEFAULT_MINSIZE,
ALLOCSET_DEFAULT_INITSIZE,
ALLOCSET_DEFAULT_MAXSIZE);
/* CDB */ /* track temp buf file allocations in separate context */
hashtable->bfCxt = AllocSetContextCreate(CurrentMemoryContext,
"hbbfcxt",
ALLOCSET_DEFAULT_MINSIZE,
ALLOCSET_DEFAULT_INITSIZE,
ALLOCSET_DEFAULT_MAXSIZE);
ExecChooseHashTableSize(outerNode->plan_rows, outerNode->plan_width,
&hashtable->nbuckets, &hashtable->nbatch, operatorMemKB);
nbuckets = hashtable->nbuckets;
nbatch = hashtable->nbatch;
hashtable->nbatch_original = nbatch;
hashtable->nbatch_outstart = nbatch;
#ifdef HJDEBUG
elog(LOG, "HJ: nbatch = %d, nbuckets = %d\n", nbatch, nbuckets);
#endif
/*
* Get info about the hash functions to be used for each hash key.
* Also remember whether the join operators are strict.
*/
nkeys = list_length(hashOperators);
hashtable->outer_hashfunctions =
(FmgrInfo *) palloc(nkeys * sizeof(FmgrInfo));
hashtable->inner_hashfunctions =
(FmgrInfo *) palloc(nkeys * sizeof(FmgrInfo));
hashtable->hashStrict = (bool *) palloc(nkeys * sizeof(bool));
i = 0;
foreach(ho, hashOperators)
{
Oid hashop = lfirst_oid(ho);
Oid left_hashfn;
Oid right_hashfn;
if (!get_op_hash_functions(hashop, &left_hashfn, &right_hashfn))
elog(ERROR, "could not find hash function for hash operator %u",
hashop);
fmgr_info(left_hashfn, &hashtable->outer_hashfunctions[i]);
fmgr_info(right_hashfn, &hashtable->inner_hashfunctions[i]);
hashtable->hashStrict[i] = op_strict(hashop);
i++;
}
/*
* Allocate data that will live for the life of the hashjoin
*/
oldcxt = MemoryContextSwitchTo(hashtable->hashCxt);
#ifdef HJDEBUG
{
/* Memory needed to allocate hashtable->batches, which consists of nbatch pointers */
int md_batch_size = (nbatch * sizeof(hashtable->batches[0])) / (1024 * 1024);
/* Memory needed to allocate hashtable->batches entries, which consist of nbatch HashJoinBatchData structures */
int md_batch_data_size = (nbatch * sizeof(HashJoinBatchData)) / (1024 * 1024);
/* Memory needed to allocate hashtable->buckets, which consists of nbuckets HashJoinTuple structures*/
int md_buckets_size = (nbuckets * sizeof(HashJoinTuple)) / (1024 * 1024);
/* Memory needed to allocate hashtable->bloom, which consists of nbuckets int64 values */
int md_bloom_size = (nbuckets * sizeof(uint64)) / (1024 * 1024);
/* Total memory needed for the hashtable metadata */
int md_tot = md_batch_size + md_batch_data_size + md_buckets_size + md_bloom_size;
elog(LOG, "About to allocate HashTable. HT_MEMORY=%dMB Memory needed for metadata: MDBATCH_ARR=%dMB, MDBATCH_DATA=%dMB, MDBUCKETS_ARR=%dMB, MDBLOOM_ARR=%dMB, TOTAL=%dMB",
(int) (hashtable->spaceAllowed / (1024 * 1024)),
md_batch_size, md_batch_data_size, md_buckets_size, md_bloom_size, md_tot);
elog(LOG, "sizeof(hashtable->batches[0])=%d, sizeof(HashJoinBatchData)=%d, sizeof(HashJoinTuple)=%d, sizeof(uint64)=%d",
(int) sizeof(hashtable->batches[0]), (int) sizeof(HashJoinBatchData),
(int) sizeof(HashJoinTuple), (int) sizeof(uint64));
}
#endif
/* array of BatchData ptrs */
hashtable->batches =
(HashJoinBatchData **)palloc(nbatch * sizeof(hashtable->batches[0]));
/* one BatchData entry per initial batch */
for (i = 0; i < nbatch; i++)
hashtable->batches[i] =
(HashJoinBatchData *)palloc0(sizeof(HashJoinBatchData));
/*
* Prepare context for the first-scan space allocations; allocate the
* hashbucket array therein, and set each bucket "empty".
*/
MemoryContextSwitchTo(hashtable->batchCxt);
hashtable->buckets = (HashJoinTuple *)
palloc0(nbuckets * sizeof(HashJoinTuple));
if(gp_hashjoin_bloomfilter!=0)
hashtable->bloom = (uint64*) palloc0(nbuckets * sizeof(uint64));
MemoryContextSwitchTo(oldcxt);
}
/* -----------------
* ExecInitGroup
*
* Creates the run-time information for the group node produced by the
* planner and initializes its outer subtree
* -----------------
*/
GroupState *
ExecInitGroup(Group *node, EState *estate, int eflags)
{
GroupState *grpstate;
/* check for unsupported flags */
Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
/*
* create state structure
*/
grpstate = makeNode(GroupState);
grpstate->ss.ps.plan = (Plan *) node;
grpstate->ss.ps.state = estate;
grpstate->grp_done = FALSE;
/*
* create expression context
*/
ExecAssignExprContext(estate, &grpstate->ss.ps);
#define GROUP_NSLOTS 2
/*
* tuple table initialization
*/
ExecInitScanTupleSlot(estate, &grpstate->ss);
ExecInitResultTupleSlot(estate, &grpstate->ss.ps);
/*
* initialize child expressions
*/
grpstate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->plan.targetlist,
(PlanState *) grpstate);
grpstate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->plan.qual,
(PlanState *) grpstate);
/*
* initialize child nodes
*/
outerPlanState(grpstate) = ExecInitNode(outerPlan(node), estate, eflags);
/*
* initialize tuple type.
*/
ExecAssignScanTypeFromOuterPlan(&grpstate->ss);
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&grpstate->ss.ps);
ExecAssignProjectionInfo(&grpstate->ss.ps, NULL);
/*
* Precompute fmgr lookup data for inner loop
*/
grpstate->eqfunctions =
execTuplesMatchPrepare(node->numCols,
node->grpOperators);
return grpstate;
}
/* ----------------------------------------------------------------
* ExecInitSort
*
* Creates the run-time state information for the sort node
* produced by the planner and initializes its outer subtree.
* ----------------------------------------------------------------
*/
SortState *
ExecInitSort(Sort *node, EState *estate, int eflags)
{
SortState *sortstate;
SO1_printf("ExecInitSort: %s\n",
"initializing sort node");
/*
* create state structure
*/
sortstate = makeNode(SortState);
sortstate->ss.ps.plan = (Plan *) node;
sortstate->ss.ps.state = estate;
/*
* We must have random access to the sort output to do backward scan or
* mark/restore. We also prefer to materialize the sort output if we
* might be called on to rewind and replay it many times.
*/
sortstate->randomAccess = (eflags & (EXEC_FLAG_REWIND |
EXEC_FLAG_BACKWARD |
EXEC_FLAG_MARK)) != 0;
sortstate->bounded = false;
sortstate->sort_Done = false;
sortstate->tuplesortstate = NULL;
/*
* Miscellaneous initialization
*
* Sort nodes don't initialize their ExprContexts because they never call
* ExecQual or ExecProject.
*/
/*
* tuple table initialization
*
* sort nodes only return scan tuples from their sorted relation.
*/
ExecInitResultTupleSlot(estate, &sortstate->ss.ps);
ExecInitScanTupleSlot(estate, &sortstate->ss);
/*
* initialize child nodes
*
* We shield the child node from the need to support REWIND, BACKWARD, or
* MARK/RESTORE.
*/
eflags &= ~(EXEC_FLAG_REWIND | EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK);
outerPlanState(sortstate) = ExecInitNode(outerPlan(node), estate, eflags);
/*
* initialize tuple type. no need to initialize projection info because
* this node doesn't do projections.
*/
ExecAssignResultTypeFromTL(&sortstate->ss.ps);
ExecAssignScanTypeFromOuterPlan(&sortstate->ss);
sortstate->ss.ps.ps_ProjInfo = NULL;
SO1_printf("ExecInitSort: %s\n",
"sort node initialized");
return sortstate;
}
/* ----------------------------------------------------------------
* ExecInitBitmapHeapScan
*
* Initializes the scan's state information.
* ----------------------------------------------------------------
*/
BitmapHeapScanState *
ExecInitBitmapHeapScan(BitmapHeapScan *node, EState *estate, int eflags)
{
BitmapHeapScanState *scanstate;
RangeTblEntry *rtentry;
Index relid;
Oid reloid;
Relation currentRelation;
/* check for unsupported flags */
Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
/*
* create state structure
*/
scanstate = makeNode(BitmapHeapScanState);
scanstate->ss.ps.plan = (Plan *) node;
scanstate->ss.ps.state = estate;
scanstate->tbm = NULL;
scanstate->tbmres = NULL;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &scanstate->ss.ps);
/*
* initialize child expressions
*/
scanstate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->scan.plan.targetlist,
(PlanState *) scanstate);
scanstate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->scan.plan.qual,
(PlanState *) scanstate);
scanstate->bitmapqualorig = (List *)
ExecInitExpr((Expr *) node->bitmapqualorig,
(PlanState *) scanstate);
#define BITMAPHEAPSCAN_NSLOTS 2
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &scanstate->ss.ps);
ExecInitScanTupleSlot(estate, &scanstate->ss);
CXT1_printf("ExecInitBitmapHeapScan: context is %d\n", CurrentMemoryContext);
/*
* open the base relation and acquire AccessShareLock on it.
*/
relid = node->scan.scanrelid;
rtentry = rt_fetch(relid, estate->es_range_table);
reloid = rtentry->relid;
currentRelation = heap_open(reloid, AccessShareLock);
scanstate->ss.ss_currentRelation = currentRelation;
/*
* Even though we aren't going to do a conventional seqscan, it is useful
* to create a HeapScanDesc --- this checks the relation size and sets up
* statistical infrastructure for us.
*/
scanstate->ss.ss_currentScanDesc = heap_beginscan(currentRelation,
estate->es_snapshot,
0,
NULL);
/*
* One problem is that heap_beginscan counts a "sequential scan" start,
* when we actually aren't doing any such thing. Reverse out the added
* scan count. (Eventually we may want to count bitmap scans separately.)
*/
pgstat_discount_heap_scan(&scanstate->ss.ss_currentScanDesc->rs_pgstat_info);
/*
* get the scan type from the relation descriptor.
*/
ExecAssignScanType(&scanstate->ss, RelationGetDescr(currentRelation), false);
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&scanstate->ss.ps);
ExecAssignScanProjectionInfo(&scanstate->ss);
/*
* initialize child nodes
*
* We do this last because the child nodes will open indexscans on our
* relation's indexes, and we want to be sure we have acquired a lock
* on the relation first.
*/
outerPlanState(scanstate) = ExecInitNode(outerPlan(node), estate, eflags);
if (IsA(outerPlan(node), BitmapIndexScan))
node->inmem =
(((BitmapIndexScan*)outerPlan(node))->indexam !=
BITMAP_AM_OID);
else if (IsA(outerPlan(node), BitmapAnd))
node->inmem = ((BitmapAnd*)outerPlan(node))->inmem;
else if (IsA(outerPlan(node), BitmapOr))
node->inmem = ((BitmapOr*)outerPlan(node))->inmem;
else
node->inmem = true;
/*
* all done.
*/
return scanstate;
}