nodeHash.h File Reference

#include "access/parallel.h"
#include "nodes/execnodes.h"

Include dependency graph for nodeHash.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Functions
HashState *	ExecInitHash (Hash *node, EState *estate, int eflags)
Node *	MultiExecHash (HashState *node)
void	ExecEndHash (HashState *node)
void	ExecReScanHash (HashState *node)
HashJoinTable	ExecHashTableCreate (HashState *state, List *hashOperators, List *hashCollations, bool keepNulls)
void	ExecParallelHashTableAlloc (HashJoinTable hashtable, int batchno)
void	ExecHashTableDestroy (HashJoinTable hashtable)
void	ExecHashTableDetach (HashJoinTable hashtable)
void	ExecHashTableDetachBatch (HashJoinTable hashtable)
void	ExecParallelHashTableSetCurrentBatch (HashJoinTable hashtable, int batchno)
void	ExecHashTableInsert (HashJoinTable hashtable, TupleTableSlot *slot, uint32 hashvalue)
void	ExecParallelHashTableInsert (HashJoinTable hashtable, TupleTableSlot *slot, uint32 hashvalue)
void	ExecParallelHashTableInsertCurrentBatch (HashJoinTable hashtable, TupleTableSlot *slot, uint32 hashvalue)
bool	ExecHashGetHashValue (HashJoinTable hashtable, ExprContext *econtext, List *hashkeys, bool outer_tuple, bool keep_nulls, uint32 *hashvalue)
void	ExecHashGetBucketAndBatch (HashJoinTable hashtable, uint32 hashvalue, int *bucketno, int *batchno)
bool	ExecScanHashBucket (HashJoinState *hjstate, ExprContext *econtext)
bool	ExecParallelScanHashBucket (HashJoinState *hjstate, ExprContext *econtext)
void	ExecPrepHashTableForUnmatched (HashJoinState *hjstate)
bool	ExecScanHashTableForUnmatched (HashJoinState *hjstate, ExprContext *econtext)
void	ExecHashTableReset (HashJoinTable hashtable)
void	ExecHashTableResetMatchFlags (HashJoinTable hashtable)
void	ExecChooseHashTableSize (double ntuples, int tupwidth, bool useskew, bool try_combined_hash_mem, int parallel_workers, size_t *space_allowed, int *numbuckets, int *numbatches, int *num_skew_mcvs)
int	ExecHashGetSkewBucket (HashJoinTable hashtable, uint32 hashvalue)
void	ExecHashEstimate (HashState *node, ParallelContext *pcxt)
void	ExecHashInitializeDSM (HashState *node, ParallelContext *pcxt)
void	ExecHashInitializeWorker (HashState *node, ParallelWorkerContext *pwcxt)
void	ExecHashRetrieveInstrumentation (HashState *node)
void	ExecShutdownHash (HashState *node)
void	ExecHashAccumInstrumentation (HashInstrumentation *instrument, HashJoinTable hashtable)

Function Documentation

ExecChooseHashTableSize()

void ExecChooseHashTableSize	(	double	ntuples,
		int	tupwidth,
		bool	useskew,
		bool	try_combined_hash_mem,
		int	parallel_workers,
		size_t *	space_allowed,
		int *	numbuckets,
		int *	numbatches,
		int *	num_skew_mcvs
	)

Definition at line 670 of file nodeHash.c.

{
    int         tupsize;
    double      inner_rel_bytes;
    size_t      hash_table_bytes;
    size_t      bucket_bytes;
    size_t      max_pointers;
    int         nbatch = 1;
    int         nbuckets;
    double      dbuckets;
 
    /* Force a plausible relation size if no info */
    if (ntuples <= 0.0)
        ntuples = 1000.0;
 
    /*
     * Estimate tupsize based on footprint of tuple in hashtable... note this
     * does not allow for any palloc overhead.  The manipulations of spaceUsed
     * don't count palloc overhead either.
     */
    tupsize = HJTUPLE_OVERHEAD +
        MAXALIGN(SizeofMinimalTupleHeader) +
        MAXALIGN(tupwidth);
    inner_rel_bytes = ntuples * tupsize;
 
    /*
     * Compute in-memory hashtable size limit from GUCs.
     */
    hash_table_bytes = get_hash_memory_limit();
 
    /*
     * Parallel Hash tries to use the combined hash_mem of all workers to
     * avoid the need to batch.  If that won't work, it falls back to hash_mem
     * per worker and tries to process batches in parallel.
     */
    if (try_combined_hash_mem)
    {
        /* Careful, this could overflow size_t */
        double      newlimit;
 
        newlimit = (double) hash_table_bytes * (double) (parallel_workers + 1);
        newlimit = Min(newlimit, (double) SIZE_MAX);
        hash_table_bytes = (size_t) newlimit;
    }
 
    *space_allowed = hash_table_bytes;
 
    /*
     * If skew optimization is possible, estimate the number of skew buckets
     * that will fit in the memory allowed, and decrement the assumed space
     * available for the main hash table accordingly.
     *
     * We make the optimistic assumption that each skew bucket will contain
     * one inner-relation tuple.  If that turns out to be low, we will recover
     * at runtime by reducing the number of skew buckets.
     *
     * hashtable->skewBucket will have up to 8 times as many HashSkewBucket
     * pointers as the number of MCVs we allow, since ExecHashBuildSkewHash
     * will round up to the next power of 2 and then multiply by 4 to reduce
     * collisions.
     */
    if (useskew)
    {
        size_t      bytes_per_mcv;
        size_t      skew_mcvs;
 
        /*----------
         * Compute number of MCVs we could hold in hash_table_bytes
         *
         * Divisor is:
         * size of a hash tuple +
         * worst-case size of skewBucket[] per MCV +
         * size of skewBucketNums[] entry +
         * size of skew bucket struct itself
         *----------
         */
        bytes_per_mcv = tupsize +
            (8 * sizeof(HashSkewBucket *)) +
            sizeof(int) +
            SKEW_BUCKET_OVERHEAD;
        skew_mcvs = hash_table_bytes / bytes_per_mcv;
 
        /*
         * Now scale by SKEW_HASH_MEM_PERCENT (we do it in this order so as
         * not to worry about size_t overflow in the multiplication)
         */
        skew_mcvs = (skew_mcvs * SKEW_HASH_MEM_PERCENT) / 100;
 
        /* Now clamp to integer range */
        skew_mcvs = Min(skew_mcvs, INT_MAX);
 
        *num_skew_mcvs = (int) skew_mcvs;
 
        /* Reduce hash_table_bytes by the amount needed for the skew table */
        if (skew_mcvs > 0)
            hash_table_bytes -= skew_mcvs * bytes_per_mcv;
    }
    else
        *num_skew_mcvs = 0;
 
    /*
     * Set nbuckets to achieve an average bucket load of NTUP_PER_BUCKET when
     * memory is filled, assuming a single batch; but limit the value so that
     * the pointer arrays we'll try to allocate do not exceed hash_table_bytes
     * nor MaxAllocSize.
     *
     * Note that both nbuckets and nbatch must be powers of 2 to make
     * ExecHashGetBucketAndBatch fast.
     */
    max_pointers = hash_table_bytes / sizeof(HashJoinTuple);
    max_pointers = Min(max_pointers, MaxAllocSize / sizeof(HashJoinTuple));
    /* If max_pointers isn't a power of 2, must round it down to one */
    max_pointers = pg_prevpower2_size_t(max_pointers);
 
    /* Also ensure we avoid integer overflow in nbatch and nbuckets */
    /* (this step is redundant given the current value of MaxAllocSize) */
    max_pointers = Min(max_pointers, INT_MAX / 2 + 1);
 
    dbuckets = ceil(ntuples / NTUP_PER_BUCKET);
    dbuckets = Min(dbuckets, max_pointers);
    nbuckets = (int) dbuckets;
    /* don't let nbuckets be really small, though ... */
    nbuckets = Max(nbuckets, 1024);
    /* ... and force it to be a power of 2. */
    nbuckets = pg_nextpower2_32(nbuckets);
 
    /*
     * If there's not enough space to store the projected number of tuples and
     * the required bucket headers, we will need multiple batches.
     */
    bucket_bytes = sizeof(HashJoinTuple) * nbuckets;
    if (inner_rel_bytes + bucket_bytes > hash_table_bytes)
    {
        /* We'll need multiple batches */
        size_t      sbuckets;
        double      dbatch;
        int         minbatch;
        size_t      bucket_size;
 
        /*
         * If Parallel Hash with combined hash_mem would still need multiple
         * batches, we'll have to fall back to regular hash_mem budget.
         */
        if (try_combined_hash_mem)
        {
            ExecChooseHashTableSize(ntuples, tupwidth, useskew,
                                    false, parallel_workers,
                                    space_allowed,
                                    numbuckets,
                                    numbatches,
                                    num_skew_mcvs);
            return;
        }
 
        /*
         * Estimate the number of buckets we'll want to have when hash_mem is
         * entirely full.  Each bucket will contain a bucket pointer plus
         * NTUP_PER_BUCKET tuples, whose projected size already includes
         * overhead for the hash code, pointer to the next tuple, etc.
         */
        bucket_size = (tupsize * NTUP_PER_BUCKET + sizeof(HashJoinTuple));
        if (hash_table_bytes <= bucket_size)
            sbuckets = 1;       /* avoid pg_nextpower2_size_t(0) */
        else
            sbuckets = pg_nextpower2_size_t(hash_table_bytes / bucket_size);
        sbuckets = Min(sbuckets, max_pointers);
        nbuckets = (int) sbuckets;
        nbuckets = pg_nextpower2_32(nbuckets);
        bucket_bytes = nbuckets * sizeof(HashJoinTuple);
 
        /*
         * Buckets are simple pointers to hashjoin tuples, while tupsize
         * includes the pointer, hash code, and MinimalTupleData.  So buckets
         * should never really exceed 25% of hash_mem (even for
         * NTUP_PER_BUCKET=1); except maybe for hash_mem values that are not
         * 2^N bytes, where we might get more because of doubling. So let's
         * look for 50% here.
         */
        Assert(bucket_bytes <= hash_table_bytes / 2);
 
        /* Calculate required number of batches. */
        dbatch = ceil(inner_rel_bytes / (hash_table_bytes - bucket_bytes));
        dbatch = Min(dbatch, max_pointers);
        minbatch = (int) dbatch;
        nbatch = pg_nextpower2_32(Max(2, minbatch));
    }
 
    Assert(nbuckets > 0);
    Assert(nbatch > 0);
 
    *numbuckets = nbuckets;
    *numbatches = nbatch;
}

References Assert(), get_hash_memory_limit(), HJTUPLE_OVERHEAD, Max, MAXALIGN, MaxAllocSize, Min, NTUP_PER_BUCKET, pg_nextpower2_32(), pg_nextpower2_size_t, pg_prevpower2_size_t, SizeofMinimalTupleHeader, SKEW_BUCKET_OVERHEAD, and SKEW_HASH_MEM_PERCENT.

Referenced by ExecHashTableCreate(), and initial_cost_hashjoin().

ExecEndHash()

void ExecEndHash

(

HashState *

node

)

Definition at line 414 of file nodeHash.c.

{
    PlanState  *outerPlan;
 
    /*
     * free exprcontext
     */
    ExecFreeExprContext(&node->ps);
 
    /*
     * shut down the subplan
     */
    outerPlan = outerPlanState(node);
    ExecEndNode(outerPlan);
}

References ExecEndNode(), ExecFreeExprContext(), outerPlan, outerPlanState, and HashState::ps.

Referenced by ExecEndNode().

ExecHashAccumInstrumentation()

void ExecHashAccumInstrumentation	(	HashInstrumentation *	instrument,
		HashJoinTable	hashtable
	)

Definition at line 2704 of file nodeHash.c.

{
    instrument->nbuckets = Max(instrument->nbuckets,
                               hashtable->nbuckets);
    instrument->nbuckets_original = Max(instrument->nbuckets_original,
                                        hashtable->nbuckets_original);
    instrument->nbatch = Max(instrument->nbatch,
                             hashtable->nbatch);
    instrument->nbatch_original = Max(instrument->nbatch_original,
                                      hashtable->nbatch_original);
    instrument->space_peak = Max(instrument->space_peak,
                                 hashtable->spacePeak);
}

References Max, HashJoinTableData::nbatch, HashInstrumentation::nbatch, HashJoinTableData::nbatch_original, HashInstrumentation::nbatch_original, HashJoinTableData::nbuckets, HashInstrumentation::nbuckets, HashJoinTableData::nbuckets_original, HashInstrumentation::nbuckets_original, HashInstrumentation::space_peak, and HashJoinTableData::spacePeak.

Referenced by ExecReScanHashJoin(), and ExecShutdownHash().

ExecHashEstimate()

void ExecHashEstimate	(	HashState *	node,
		ParallelContext *	pcxt
	)

Definition at line 2588 of file nodeHash.c.

{
    size_t      size;
 
    /* don't need this if not instrumenting or no workers */
    if (!node->ps.instrument || pcxt->nworkers == 0)
        return;
 
    size = mul_size(pcxt->nworkers, sizeof(HashInstrumentation));
    size = add_size(size, offsetof(SharedHashInfo, hinstrument));
    shm_toc_estimate_chunk(&pcxt->estimator, size);
    shm_toc_estimate_keys(&pcxt->estimator, 1);
}

References add_size(), ParallelContext::estimator, PlanState::instrument, mul_size(), ParallelContext::nworkers, HashState::ps, shm_toc_estimate_chunk, and shm_toc_estimate_keys.

Referenced by ExecParallelEstimate().

ExecHashGetBucketAndBatch()

void ExecHashGetBucketAndBatch	(	HashJoinTable	hashtable,
		uint32	hashvalue,
		int *	bucketno,
		int *	batchno
	)

Definition at line 1915 of file nodeHash.c.

{
    uint32      nbuckets = (uint32) hashtable->nbuckets;
    uint32      nbatch = (uint32) hashtable->nbatch;
 
    if (nbatch > 1)
    {
        *bucketno = hashvalue & (nbuckets - 1);
        *batchno = pg_rotate_right32(hashvalue,
                                     hashtable->log2_nbuckets) & (nbatch - 1);
    }
    else
    {
        *bucketno = hashvalue & (nbuckets - 1);
        *batchno = 0;
    }
}

References HashJoinTableData::log2_nbuckets, HashJoinTableData::nbatch, HashJoinTableData::nbuckets, and pg_rotate_right32().

Referenced by ExecHashIncreaseNumBatches(), ExecHashIncreaseNumBuckets(), ExecHashJoinImpl(), ExecHashRemoveNextSkewBucket(), ExecHashTableInsert(), ExecParallelHashIncreaseNumBuckets(), ExecParallelHashJoinPartitionOuter(), ExecParallelHashRepartitionFirst(), ExecParallelHashRepartitionRest(), ExecParallelHashTableInsert(), and ExecParallelHashTableInsertCurrentBatch().

ExecHashGetHashValue()

bool ExecHashGetHashValue	(	HashJoinTable	hashtable,
		ExprContext *	econtext,
		List *	hashkeys,
		bool	outer_tuple,
		bool	keep_nulls,
		uint32 *	hashvalue
	)

Definition at line 1807 of file nodeHash.c.

{
    uint32      hashkey = 0;
    FmgrInfo   *hashfunctions;
    ListCell   *hk;
    int         i = 0;
    MemoryContext oldContext;
 
    /*
     * We reset the eval context each time to reclaim any memory leaked in the
     * hashkey expressions.
     */
    ResetExprContext(econtext);
 
    oldContext = MemoryContextSwitchTo(econtext->ecxt_per_tuple_memory);
 
    if (outer_tuple)
        hashfunctions = hashtable->outer_hashfunctions;
    else
        hashfunctions = hashtable->inner_hashfunctions;
 
    foreach(hk, hashkeys)
    {
        ExprState  *keyexpr = (ExprState *) lfirst(hk);
        Datum       keyval;
        bool        isNull;
 
        /* combine successive hashkeys by rotating */
        hashkey = pg_rotate_left32(hashkey, 1);
 
        /*
         * Get the join attribute value of the tuple
         */
        keyval = ExecEvalExpr(keyexpr, econtext, &isNull);
 
        /*
         * If the attribute is NULL, and the join operator is strict, then
         * this tuple cannot pass the join qual so we can reject it
         * immediately (unless we're scanning the outside of an outer join, in
         * which case we must not reject it).  Otherwise we act like the
         * hashcode of NULL is zero (this will support operators that act like
         * IS NOT DISTINCT, though not any more-random behavior).  We treat
         * the hash support function as strict even if the operator is not.
         *
         * Note: currently, all hashjoinable operators must be strict since
         * the hash index AM assumes that.  However, it takes so little extra
         * code here to allow non-strict that we may as well do it.
         */
        if (isNull)
        {
            if (hashtable->hashStrict[i] && !keep_nulls)
            {
                MemoryContextSwitchTo(oldContext);
                return false;   /* cannot match */
            }
            /* else, leave hashkey unmodified, equivalent to hashcode 0 */
        }
        else
        {
            /* Compute the hash function */
            uint32      hkey;
 
            hkey = DatumGetUInt32(FunctionCall1Coll(&hashfunctions[i], hashtable->collations[i], keyval));
            hashkey ^= hkey;
        }
 
        i++;
    }
 
    MemoryContextSwitchTo(oldContext);
 
    *hashvalue = hashkey;
    return true;
}

References HashJoinTableData::collations, DatumGetUInt32(), ExprContext::ecxt_per_tuple_memory, ExecEvalExpr(), FunctionCall1Coll(), HashJoinTableData::hashStrict, i, HashJoinTableData::inner_hashfunctions, lfirst, MemoryContextSwitchTo(), HashJoinTableData::outer_hashfunctions, pg_rotate_left32(), and ResetExprContext.

Referenced by ExecHashJoinOuterGetTuple(), ExecParallelHashJoinOuterGetTuple(), ExecParallelHashJoinPartitionOuter(), MultiExecParallelHash(), and MultiExecPrivateHash().

ExecHashGetSkewBucket()

int ExecHashGetSkewBucket	(	HashJoinTable	hashtable,
		uint32	hashvalue
	)

Definition at line 2383 of file nodeHash.c.

{
    int         bucket;
 
    /*
     * Always return INVALID_SKEW_BUCKET_NO if not doing skew optimization (in
     * particular, this happens after the initial batch is done).
     */
    if (!hashtable->skewEnabled)
        return INVALID_SKEW_BUCKET_NO;
 
    /*
     * Since skewBucketLen is a power of 2, we can do a modulo by ANDing.
     */
    bucket = hashvalue & (hashtable->skewBucketLen - 1);
 
    /*
     * While we have not hit a hole in the hashtable and have not hit the
     * desired bucket, we have collided with some other hash value, so try the
     * next bucket location.
     */
    while (hashtable->skewBucket[bucket] != NULL &&
           hashtable->skewBucket[bucket]->hashvalue != hashvalue)
        bucket = (bucket + 1) & (hashtable->skewBucketLen - 1);
 
    /*
     * Found the desired bucket?
     */
    if (hashtable->skewBucket[bucket] != NULL)
        return bucket;
 
    /*
     * There must not be any hashtable entry for this hash value.
     */
    return INVALID_SKEW_BUCKET_NO;
}

References HashSkewBucket::hashvalue, INVALID_SKEW_BUCKET_NO, HashJoinTableData::skewBucket, HashJoinTableData::skewBucketLen, and HashJoinTableData::skewEnabled.

Referenced by ExecHashJoinImpl(), and MultiExecPrivateHash().

ExecHashInitializeDSM()

void ExecHashInitializeDSM	(	HashState *	node,
		ParallelContext *	pcxt
	)

Definition at line 2607 of file nodeHash.c.

{
    size_t      size;
 
    /* don't need this if not instrumenting or no workers */
    if (!node->ps.instrument || pcxt->nworkers == 0)
        return;
 
    size = offsetof(SharedHashInfo, hinstrument) +
        pcxt->nworkers * sizeof(HashInstrumentation);
    node->shared_info = (SharedHashInfo *) shm_toc_allocate(pcxt->toc, size);
 
    /* Each per-worker area must start out as zeroes. */
    memset(node->shared_info, 0, size);
 
    node->shared_info->num_workers = pcxt->nworkers;
    shm_toc_insert(pcxt->toc, node->ps.plan->plan_node_id,
                   node->shared_info);
}

References PlanState::instrument, SharedHashInfo::num_workers, ParallelContext::nworkers, PlanState::plan, Plan::plan_node_id, HashState::ps, HashState::shared_info, shm_toc_allocate(), shm_toc_insert(), and ParallelContext::toc.

Referenced by ExecParallelInitializeDSM().

ExecHashInitializeWorker()

void ExecHashInitializeWorker	(	HashState *	node,
		ParallelWorkerContext *	pwcxt
	)

Definition at line 2632 of file nodeHash.c.

{
    SharedHashInfo *shared_info;
 
    /* don't need this if not instrumenting */
    if (!node->ps.instrument)
        return;
 
    /*
     * Find our entry in the shared area, and set up a pointer to it so that
     * we'll accumulate stats there when shutting down or rebuilding the hash
     * table.
     */
    shared_info = (SharedHashInfo *)
        shm_toc_lookup(pwcxt->toc, node->ps.plan->plan_node_id, false);
    node->hinstrument = &shared_info->hinstrument[ParallelWorkerNumber];
}

References SharedHashInfo::hinstrument, HashState::hinstrument, PlanState::instrument, ParallelWorkerNumber, PlanState::plan, Plan::plan_node_id, HashState::ps, shm_toc_lookup(), and ParallelWorkerContext::toc.

Referenced by ExecParallelInitializeWorker().

ExecHashRetrieveInstrumentation()

void ExecHashRetrieveInstrumentation

(

HashState *

node

)

Definition at line 2673 of file nodeHash.c.

{
    SharedHashInfo *shared_info = node->shared_info;
    size_t      size;
 
    if (shared_info == NULL)
        return;
 
    /* Replace node->shared_info with a copy in backend-local memory. */
    size = offsetof(SharedHashInfo, hinstrument) +
        shared_info->num_workers * sizeof(HashInstrumentation);
    node->shared_info = palloc(size);
    memcpy(node->shared_info, shared_info, size);
}

References SharedHashInfo::num_workers, palloc(), and HashState::shared_info.

Referenced by ExecParallelRetrieveInstrumentation().

ExecHashTableCreate()

HashJoinTable ExecHashTableCreate	(	HashState *	state,
		List *	hashOperators,
		List *	hashCollations,
		bool	keepNulls
	)

Definition at line 438 of file nodeHash.c.

{
    Hash       *node;
    HashJoinTable hashtable;
    Plan       *outerNode;
    size_t      space_allowed;
    int         nbuckets;
    int         nbatch;
    double      rows;
    int         num_skew_mcvs;
    int         log2_nbuckets;
    int         nkeys;
    int         i;
    ListCell   *ho;
    ListCell   *hc;
    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.
     */
    node = (Hash *) state->ps.plan;
    outerNode = outerPlan(node);
 
    /*
     * If this is shared hash table with a partial plan, then we can't use
     * outerNode->plan_rows to estimate its size.  We need an estimate of the
     * total number of rows across all copies of the partial plan.
     */
    rows = node->plan.parallel_aware ? node->rows_total : outerNode->plan_rows;
 
    ExecChooseHashTableSize(rows, outerNode->plan_width,
                            OidIsValid(node->skewTable),
                            state->parallel_state != NULL,
                            state->parallel_state != NULL ?
                            state->parallel_state->nparticipants - 1 : 0,
                            &space_allowed,
                            &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 = palloc_object(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.unshared = 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->partialTuples = 0;
    hashtable->skewTuples = 0;
    hashtable->innerBatchFile = NULL;
    hashtable->outerBatchFile = NULL;
    hashtable->spaceUsed = 0;
    hashtable->spacePeak = 0;
    hashtable->spaceAllowed = space_allowed;
    hashtable->spaceUsedSkew = 0;
    hashtable->spaceAllowedSkew =
        hashtable->spaceAllowed * SKEW_HASH_MEM_PERCENT / 100;
    hashtable->chunks = NULL;
    hashtable->current_chunk = NULL;
    hashtable->parallel_state = state->parallel_state;
    hashtable->area = state->ps.state->es_query_dsa;
    hashtable->batches = 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 = palloc_array(FmgrInfo, nkeys);
    hashtable->inner_hashfunctions = palloc_array(FmgrInfo, nkeys);
    hashtable->hashStrict = palloc_array(bool, nkeys);
    hashtable->collations = palloc_array(Oid, nkeys);
    i = 0;
    forboth(ho, hashOperators, hc, hashCollations)
    {
        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);
        hashtable->collations[i] = lfirst_oid(hc);
        i++;
    }
 
    if (nbatch > 1 && hashtable->parallel_state == NULL)
    {
        /*
         * allocate and initialize the file arrays in hashCxt (not needed for
         * parallel case which uses shared tuplestores instead of raw files)
         */
        hashtable->innerBatchFile = palloc0_array(BufFile *, nbatch);
        hashtable->outerBatchFile = palloc0_array(BufFile *, nbatch);
        /* The files will not be opened until needed... */
        /* ... but make sure we have temp tablespaces established for them */
        PrepareTempTablespaces();
    }
 
    MemoryContextSwitchTo(oldcxt);
 
    if (hashtable->parallel_state)
    {
        ParallelHashJoinState *pstate = hashtable->parallel_state;
        Barrier    *build_barrier;
 
        /*
         * Attach to the build barrier.  The corresponding detach operation is
         * in ExecHashTableDetach.  Note that we won't attach to the
         * batch_barrier for batch 0 yet.  We'll attach later and start it out
         * in PHJ_BATCH_PROBE phase, because batch 0 is allocated up front and
         * then loaded while hashing (the standard hybrid hash join
         * algorithm), and we'll coordinate that using build_barrier.
         */
        build_barrier = &pstate->build_barrier;
        BarrierAttach(build_barrier);
 
        /*
         * So far we have no idea whether there are any other participants,
         * and if so, what phase they are working on.  The only thing we care
         * about at this point is whether someone has already created the
         * SharedHashJoinBatch objects and the hash table for batch 0.  One
         * backend will be elected to do that now if necessary.
         */
        if (BarrierPhase(build_barrier) == PHJ_BUILD_ELECT &&
            BarrierArriveAndWait(build_barrier, WAIT_EVENT_HASH_BUILD_ELECT))
        {
            pstate->nbatch = nbatch;
            pstate->space_allowed = space_allowed;
            pstate->growth = PHJ_GROWTH_OK;
 
            /* Set up the shared state for coordinating batches. */
            ExecParallelHashJoinSetUpBatches(hashtable, nbatch);
 
            /*
             * Allocate batch 0's hash table up front so we can load it
             * directly while hashing.
             */
            pstate->nbuckets = nbuckets;
            ExecParallelHashTableAlloc(hashtable, 0);
        }
 
        /*
         * The next Parallel Hash synchronization point is in
         * MultiExecParallelHash(), which will progress it all the way to
         * PHJ_BUILD_RUN.  The caller must not return control from this
         * executor node between now and then.
         */
    }
    else
    {
        /*
         * Prepare context for the first-scan space allocations; allocate the
         * hashbucket array therein, and set each bucket "empty".
         */
        MemoryContextSwitchTo(hashtable->batchCxt);
 
        hashtable->buckets.unshared = palloc0_array(HashJoinTuple, nbuckets);
 
        /*
         * Set up for skew optimization, if possible and there's a need for
         * more than one batch.  (In a one-batch join, there's no point in
         * it.)
         */
        if (nbatch > 1)
            ExecHashBuildSkewHash(hashtable, node, num_skew_mcvs);
 
        MemoryContextSwitchTo(oldcxt);
    }
 
    return hashtable;
}

References ALLOCSET_DEFAULT_SIZES, AllocSetContextCreate, HashJoinTableData::area, Assert(), BarrierArriveAndWait(), BarrierAttach(), BarrierPhase(), HashJoinTableData::batchCxt, HashJoinTableData::batches, HashJoinTableData::buckets, ParallelHashJoinState::build_barrier, HashJoinTableData::chunks, HashJoinTableData::collations, HashJoinTableData::curbatch, HashJoinTableData::current_chunk, CurrentMemoryContext, elog(), ERROR, ExecChooseHashTableSize(), ExecHashBuildSkewHash(), ExecParallelHashJoinSetUpBatches(), ExecParallelHashTableAlloc(), fmgr_info(), forboth, get_op_hash_functions(), HashJoinTableData::growEnabled, ParallelHashJoinState::growth, HashJoinTableData::hashCxt, HashJoinTableData::hashStrict, i, HashJoinTableData::inner_hashfunctions, HashJoinTableData::innerBatchFile, HashJoinTableData::keepNulls, lfirst_oid, list_length(), HashJoinTableData::log2_nbuckets, HashJoinTableData::log2_nbuckets_optimal, MemoryContextSwitchTo(), my_log2(), ParallelHashJoinState::nbatch, HashJoinTableData::nbatch, HashJoinTableData::nbatch_original, HashJoinTableData::nbatch_outstart, ParallelHashJoinState::nbuckets, HashJoinTableData::nbuckets, HashJoinTableData::nbuckets_optimal, HashJoinTableData::nbuckets_original, HashJoinTableData::nSkewBuckets, OidIsValid, op_strict(), HashJoinTableData::outer_hashfunctions, HashJoinTableData::outerBatchFile, outerPlan, palloc0_array, palloc_array, palloc_object, Plan::parallel_aware, HashJoinTableData::parallel_state, HashJoinTableData::partialTuples, PHJ_BUILD_ELECT, PHJ_GROWTH_OK, Hash::plan, Plan::plan_rows, Plan::plan_width, PrepareTempTablespaces(), printf, Hash::rows_total, SKEW_HASH_MEM_PERCENT, HashJoinTableData::skewBucket, HashJoinTableData::skewBucketLen, HashJoinTableData::skewBucketNums, HashJoinTableData::skewEnabled, Hash::skewTable, HashJoinTableData::skewTuples, ParallelHashJoinState::space_allowed, HashJoinTableData::spaceAllowed, HashJoinTableData::spaceAllowedSkew, HashJoinTableData::spacePeak, HashJoinTableData::spaceUsed, HashJoinTableData::spaceUsedSkew, HashJoinTableData::totalTuples, HashJoinTableData::unshared, and WAIT_EVENT_HASH_BUILD_ELECT.

Referenced by ExecHashJoinImpl().

ExecHashTableDestroy()

void ExecHashTableDestroy

(

HashJoinTable

hashtable

)

Definition at line 878 of file nodeHash.c.

{
    int         i;
 
    /*
     * Make sure all the temp files are closed.  We skip batch 0, since it
     * can't have any temp files (and the arrays might not even exist if
     * nbatch is only 1).  Parallel hash joins don't use these files.
     */
    if (hashtable->innerBatchFile != NULL)
    {
        for (i = 1; i < hashtable->nbatch; i++)
        {
            if (hashtable->innerBatchFile[i])
                BufFileClose(hashtable->innerBatchFile[i]);
            if (hashtable->outerBatchFile[i])
                BufFileClose(hashtable->outerBatchFile[i]);
        }
    }
 
    /* Release working memory (batchCxt is a child, so it goes away too) */
    MemoryContextDelete(hashtable->hashCxt);
 
    /* And drop the control block */
    pfree(hashtable);
}

References BufFileClose(), HashJoinTableData::hashCxt, i, HashJoinTableData::innerBatchFile, MemoryContextDelete(), HashJoinTableData::nbatch, HashJoinTableData::outerBatchFile, and pfree().

Referenced by ExecEndHashJoin(), and ExecReScanHashJoin().

ExecHashTableDetach()

void ExecHashTableDetach

(

HashJoinTable

hashtable

)

Definition at line 3186 of file nodeHash.c.

{
    ParallelHashJoinState *pstate = hashtable->parallel_state;
 
    /*
     * If we're involved in a parallel query, we must either have gotten all
     * the way to PHJ_BUILD_RUN, or joined too late and be in PHJ_BUILD_FREE.
     */
    Assert(!pstate ||
           BarrierPhase(&pstate->build_barrier) >= PHJ_BUILD_RUN);
 
    if (pstate && BarrierPhase(&pstate->build_barrier) == PHJ_BUILD_RUN)
    {
        int         i;
 
        /* Make sure any temporary files are closed. */
        if (hashtable->batches)
        {
            for (i = 0; i < hashtable->nbatch; ++i)
            {
                sts_end_write(hashtable->batches[i].inner_tuples);
                sts_end_write(hashtable->batches[i].outer_tuples);
                sts_end_parallel_scan(hashtable->batches[i].inner_tuples);
                sts_end_parallel_scan(hashtable->batches[i].outer_tuples);
            }
        }
 
        /* If we're last to detach, clean up shared memory. */
        if (BarrierArriveAndDetach(&pstate->build_barrier))
        {
            /*
             * Late joining processes will see this state and give up
             * immediately.
             */
            Assert(BarrierPhase(&pstate->build_barrier) == PHJ_BUILD_FREE);
 
            if (DsaPointerIsValid(pstate->batches))
            {
                dsa_free(hashtable->area, pstate->batches);
                pstate->batches = InvalidDsaPointer;
            }
        }
    }
    hashtable->parallel_state = NULL;
}

References HashJoinTableData::area, Assert(), BarrierArriveAndDetach(), BarrierPhase(), ParallelHashJoinState::batches, HashJoinTableData::batches, ParallelHashJoinState::build_barrier, dsa_free(), DsaPointerIsValid, i, ParallelHashJoinBatchAccessor::inner_tuples, InvalidDsaPointer, HashJoinTableData::nbatch, ParallelHashJoinBatchAccessor::outer_tuples, HashJoinTableData::parallel_state, PHJ_BUILD_FREE, PHJ_BUILD_RUN, sts_end_parallel_scan(), and sts_end_write().

Referenced by ExecHashJoinReInitializeDSM(), and ExecShutdownHashJoin().

ExecHashTableDetachBatch()

void ExecHashTableDetachBatch

(

HashJoinTable

hashtable

)

Definition at line 3129 of file nodeHash.c.

{
    if (hashtable->parallel_state != NULL &&
        hashtable->curbatch >= 0)
    {
        int         curbatch = hashtable->curbatch;
        ParallelHashJoinBatch *batch = hashtable->batches[curbatch].shared;
 
        /* Make sure any temporary files are closed. */
        sts_end_parallel_scan(hashtable->batches[curbatch].inner_tuples);
        sts_end_parallel_scan(hashtable->batches[curbatch].outer_tuples);
 
        /* Detach from the batch we were last working on. */
        if (BarrierArriveAndDetach(&batch->batch_barrier))
        {
            /*
             * Technically we shouldn't access the barrier because we're no
             * longer attached, but since there is no way it's moving after
             * this point it seems safe to make the following assertion.
             */
            Assert(BarrierPhase(&batch->batch_barrier) == PHJ_BATCH_FREE);
 
            /* Free shared chunks and buckets. */
            while (DsaPointerIsValid(batch->chunks))
            {
                HashMemoryChunk chunk =
                dsa_get_address(hashtable->area, batch->chunks);
                dsa_pointer next = chunk->next.shared;
 
                dsa_free(hashtable->area, batch->chunks);
                batch->chunks = next;
            }
            if (DsaPointerIsValid(batch->buckets))
            {
                dsa_free(hashtable->area, batch->buckets);
                batch->buckets = InvalidDsaPointer;
            }
        }
 
        /*
         * Track the largest batch we've been attached to.  Though each
         * backend might see a different subset of batches, explain.c will
         * scan the results from all backends to find the largest value.
         */
        hashtable->spacePeak =
            Max(hashtable->spacePeak,
                batch->size + sizeof(dsa_pointer_atomic) * hashtable->nbuckets);
 
        /* Remember that we are not attached to a batch. */
        hashtable->curbatch = -1;
    }
}

References HashJoinTableData::area, Assert(), BarrierArriveAndDetach(), BarrierPhase(), ParallelHashJoinBatch::batch_barrier, HashJoinTableData::batches, ParallelHashJoinBatch::buckets, ParallelHashJoinBatch::chunks, HashJoinTableData::curbatch, dsa_free(), dsa_get_address(), DsaPointerIsValid, ParallelHashJoinBatchAccessor::inner_tuples, InvalidDsaPointer, Max, HashJoinTableData::nbuckets, next, HashMemoryChunkData::next, ParallelHashJoinBatchAccessor::outer_tuples, HashJoinTableData::parallel_state, PHJ_BATCH_FREE, HashMemoryChunkData::shared, ParallelHashJoinBatchAccessor::shared, ParallelHashJoinBatch::size, HashJoinTableData::spacePeak, and sts_end_parallel_scan().

Referenced by ExecHashJoinReInitializeDSM(), ExecParallelHashJoinNewBatch(), and ExecShutdownHashJoin().

ExecHashTableInsert()

void ExecHashTableInsert	(	HashJoinTable	hashtable,
		TupleTableSlot *	slot,
		uint32	hashvalue
	)

Definition at line 1609 of file nodeHash.c.

{
    bool        shouldFree;
    MinimalTuple tuple = ExecFetchSlotMinimalTuple(slot, &shouldFree);
    int         bucketno;
    int         batchno;
 
    ExecHashGetBucketAndBatch(hashtable, hashvalue,
                              &bucketno, &batchno);
 
    /*
     * decide whether to put the tuple in the hash table or a temp file
     */
    if (batchno == hashtable->curbatch)
    {
        /*
         * put the tuple in hash table
         */
        HashJoinTuple hashTuple;
        int         hashTupleSize;
        double      ntuples = (hashtable->totalTuples - hashtable->skewTuples);
 
        /* Create the HashJoinTuple */
        hashTupleSize = HJTUPLE_OVERHEAD + tuple->t_len;
        hashTuple = (HashJoinTuple) dense_alloc(hashtable, hashTupleSize);
 
        hashTuple->hashvalue = hashvalue;
        memcpy(HJTUPLE_MINTUPLE(hashTuple), tuple, tuple->t_len);
 
        /*
         * We always reset the tuple-matched flag on insertion.  This is okay
         * even when reloading a tuple from a batch file, since the tuple
         * could not possibly have been matched to an outer tuple before it
         * went into the batch file.
         */
        HeapTupleHeaderClearMatch(HJTUPLE_MINTUPLE(hashTuple));
 
        /* Push it onto the front of the bucket's list */
        hashTuple->next.unshared = hashtable->buckets.unshared[bucketno];
        hashtable->buckets.unshared[bucketno] = hashTuple;
 
        /*
         * Increase the (optimal) number of buckets if we just exceeded the
         * NTUP_PER_BUCKET threshold, but only when there's still a single
         * batch.
         */
        if (hashtable->nbatch == 1 &&
            ntuples > (hashtable->nbuckets_optimal * NTUP_PER_BUCKET))
        {
            /* Guard against integer overflow and alloc size overflow */
            if (hashtable->nbuckets_optimal <= INT_MAX / 2 &&
                hashtable->nbuckets_optimal * 2 <= MaxAllocSize / sizeof(HashJoinTuple))
            {
                hashtable->nbuckets_optimal *= 2;
                hashtable->log2_nbuckets_optimal += 1;
            }
        }
 
        /* Account for space used, and back off if we've used too much */
        hashtable->spaceUsed += hashTupleSize;
        if (hashtable->spaceUsed > hashtable->spacePeak)
            hashtable->spacePeak = hashtable->spaceUsed;
        if (hashtable->spaceUsed +
            hashtable->nbuckets_optimal * sizeof(HashJoinTuple)
            > hashtable->spaceAllowed)
            ExecHashIncreaseNumBatches(hashtable);
    }
    else
    {
        /*
         * put the tuple into a temp file for later batches
         */
        Assert(batchno > hashtable->curbatch);
        ExecHashJoinSaveTuple(tuple,
                              hashvalue,
                              &hashtable->innerBatchFile[batchno]);
    }
 
    if (shouldFree)
        heap_free_minimal_tuple(tuple);
}

References Assert(), HashJoinTableData::buckets, HashJoinTableData::curbatch, dense_alloc(), ExecFetchSlotMinimalTuple(), ExecHashGetBucketAndBatch(), ExecHashIncreaseNumBatches(), ExecHashJoinSaveTuple(), HashJoinTupleData::hashvalue, heap_free_minimal_tuple(), HeapTupleHeaderClearMatch, HJTUPLE_MINTUPLE, HJTUPLE_OVERHEAD, HashJoinTableData::innerBatchFile, HashJoinTableData::log2_nbuckets_optimal, MaxAllocSize, HashJoinTableData::nbatch, HashJoinTableData::nbuckets_optimal, HashJoinTupleData::next, NTUP_PER_BUCKET, HashJoinTableData::skewTuples, HashJoinTableData::spaceAllowed, HashJoinTableData::spacePeak, HashJoinTableData::spaceUsed, MinimalTupleData::t_len, HashJoinTableData::totalTuples, HashJoinTupleData::unshared, and HashJoinTableData::unshared.

Referenced by ExecHashJoinNewBatch(), and MultiExecPrivateHash().

ExecHashTableReset()

void ExecHashTableReset

(

HashJoinTable

hashtable

)

Definition at line 2154 of file nodeHash.c.

{
    MemoryContext oldcxt;
    int         nbuckets = hashtable->nbuckets;
 
    /*
     * Release all the hash buckets and tuples acquired in the prior pass, and
     * reinitialize the context for a new pass.
     */
    MemoryContextReset(hashtable->batchCxt);
    oldcxt = MemoryContextSwitchTo(hashtable->batchCxt);
 
    /* Reallocate and reinitialize the hash bucket headers. */
    hashtable->buckets.unshared = palloc0_array(HashJoinTuple, nbuckets);
 
    hashtable->spaceUsed = 0;
 
    MemoryContextSwitchTo(oldcxt);
 
    /* Forget the chunks (the memory was freed by the context reset above). */
    hashtable->chunks = NULL;
}

References HashJoinTableData::batchCxt, HashJoinTableData::buckets, HashJoinTableData::chunks, MemoryContextReset(), MemoryContextSwitchTo(), HashJoinTableData::nbuckets, palloc0_array, HashJoinTableData::spaceUsed, and HashJoinTableData::unshared.

Referenced by ExecHashJoinNewBatch().

ExecHashTableResetMatchFlags()

void ExecHashTableResetMatchFlags

(

HashJoinTable

hashtable

)

Definition at line 2182 of file nodeHash.c.

{
    HashJoinTuple tuple;
    int         i;
 
    /* Reset all flags in the main table ... */
    for (i = 0; i < hashtable->nbuckets; i++)
    {
        for (tuple = hashtable->buckets.unshared[i]; tuple != NULL;
             tuple = tuple->next.unshared)
            HeapTupleHeaderClearMatch(HJTUPLE_MINTUPLE(tuple));
    }
 
    /* ... and the same for the skew buckets, if any */
    for (i = 0; i < hashtable->nSkewBuckets; i++)
    {
        int         j = hashtable->skewBucketNums[i];
        HashSkewBucket *skewBucket = hashtable->skewBucket[j];
 
        for (tuple = skewBucket->tuples; tuple != NULL; tuple = tuple->next.unshared)
            HeapTupleHeaderClearMatch(HJTUPLE_MINTUPLE(tuple));
    }
}

References HashJoinTableData::buckets, HeapTupleHeaderClearMatch, HJTUPLE_MINTUPLE, i, j, HashJoinTableData::nbuckets, HashJoinTupleData::next, HashJoinTableData::nSkewBuckets, HashJoinTableData::skewBucket, HashJoinTableData::skewBucketNums, HashSkewBucket::tuples, HashJoinTupleData::unshared, and HashJoinTableData::unshared.

Referenced by ExecReScanHashJoin().

ExecInitHash()

HashState* ExecInitHash	(	Hash *	node,
		EState *	estate,
		int	eflags
	)

Definition at line 361 of file nodeHash.c.

{
    HashState  *hashstate;
 
    /* check for unsupported flags */
    Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
 
    /*
     * create state structure
     */
    hashstate = makeNode(HashState);
    hashstate->ps.plan = (Plan *) node;
    hashstate->ps.state = estate;
    hashstate->ps.ExecProcNode = ExecHash;
    hashstate->hashtable = NULL;
    hashstate->hashkeys = NIL;  /* will be set by parent HashJoin */
 
    /*
     * Miscellaneous initialization
     *
     * create expression context for node
     */
    ExecAssignExprContext(estate, &hashstate->ps);
 
    /*
     * initialize child nodes
     */
    outerPlanState(hashstate) = ExecInitNode(outerPlan(node), estate, eflags);
 
    /*
     * initialize our result slot and type. No need to build projection
     * because this node doesn't do projections.
     */
    ExecInitResultTupleSlotTL(&hashstate->ps, &TTSOpsMinimalTuple);
    hashstate->ps.ps_ProjInfo = NULL;
 
    /*
     * initialize child expressions
     */
    Assert(node->plan.qual == NIL);
    hashstate->hashkeys =
        ExecInitExprList(node->hashkeys, (PlanState *) hashstate);
 
    return hashstate;
}

References Assert(), EXEC_FLAG_BACKWARD, EXEC_FLAG_MARK, ExecAssignExprContext(), ExecHash(), ExecInitExprList(), ExecInitNode(), ExecInitResultTupleSlotTL(), PlanState::ExecProcNode, HashState::hashkeys, Hash::hashkeys, HashState::hashtable, makeNode, NIL, outerPlan, outerPlanState, PlanState::plan, Hash::plan, HashState::ps, PlanState::ps_ProjInfo, Plan::qual, PlanState::state, and TTSOpsMinimalTuple.

Referenced by ExecInitNode().

ExecParallelHashTableAlloc()

void ExecParallelHashTableAlloc	(	HashJoinTable	hashtable,
		int	batchno
	)

Definition at line 3109 of file nodeHash.c.

{
    ParallelHashJoinBatch *batch = hashtable->batches[batchno].shared;
    dsa_pointer_atomic *buckets;
    int         nbuckets = hashtable->parallel_state->nbuckets;
    int         i;
 
    batch->buckets =
        dsa_allocate(hashtable->area, sizeof(dsa_pointer_atomic) * nbuckets);
    buckets = (dsa_pointer_atomic *)
        dsa_get_address(hashtable->area, batch->buckets);
    for (i = 0; i < nbuckets; ++i)
        dsa_pointer_atomic_init(&buckets[i], InvalidDsaPointer);
}

References HashJoinTableData::area, HashJoinTableData::batches, ParallelHashJoinBatch::buckets, dsa_allocate, dsa_get_address(), dsa_pointer_atomic_init, i, InvalidDsaPointer, ParallelHashJoinState::nbuckets, HashJoinTableData::parallel_state, and ParallelHashJoinBatchAccessor::shared.

Referenced by ExecHashTableCreate(), and ExecParallelHashJoinNewBatch().

ExecParallelHashTableInsert()

void ExecParallelHashTableInsert	(	HashJoinTable	hashtable,
		TupleTableSlot *	slot,
		uint32	hashvalue
	)

Definition at line 1698 of file nodeHash.c.

{
    bool        shouldFree;
    MinimalTuple tuple = ExecFetchSlotMinimalTuple(slot, &shouldFree);
    dsa_pointer shared;
    int         bucketno;
    int         batchno;
 
retry:
    ExecHashGetBucketAndBatch(hashtable, hashvalue, &bucketno, &batchno);
 
    if (batchno == 0)
    {
        HashJoinTuple hashTuple;
 
        /* Try to load it into memory. */
        Assert(BarrierPhase(&hashtable->parallel_state->build_barrier) ==
               PHJ_BUILD_HASH_INNER);
        hashTuple = ExecParallelHashTupleAlloc(hashtable,
                                               HJTUPLE_OVERHEAD + tuple->t_len,
                                               &shared);
        if (hashTuple == NULL)
            goto retry;
 
        /* Store the hash value in the HashJoinTuple header. */
        hashTuple->hashvalue = hashvalue;
        memcpy(HJTUPLE_MINTUPLE(hashTuple), tuple, tuple->t_len);
 
        /* Push it onto the front of the bucket's list */
        ExecParallelHashPushTuple(&hashtable->buckets.shared[bucketno],
                                  hashTuple, shared);
    }
    else
    {
        size_t      tuple_size = MAXALIGN(HJTUPLE_OVERHEAD + tuple->t_len);
 
        Assert(batchno > 0);
 
        /* Try to preallocate space in the batch if necessary. */
        if (hashtable->batches[batchno].preallocated < tuple_size)
        {
            if (!ExecParallelHashTuplePrealloc(hashtable, batchno, tuple_size))
                goto retry;
        }
 
        Assert(hashtable->batches[batchno].preallocated >= tuple_size);
        hashtable->batches[batchno].preallocated -= tuple_size;
        sts_puttuple(hashtable->batches[batchno].inner_tuples, &hashvalue,
                     tuple);
    }
    ++hashtable->batches[batchno].ntuples;
 
    if (shouldFree)
        heap_free_minimal_tuple(tuple);
}

References Assert(), BarrierPhase(), HashJoinTableData::batches, HashJoinTableData::buckets, ParallelHashJoinState::build_barrier, ExecFetchSlotMinimalTuple(), ExecHashGetBucketAndBatch(), ExecParallelHashPushTuple(), ExecParallelHashTupleAlloc(), ExecParallelHashTuplePrealloc(), HashJoinTupleData::hashvalue, heap_free_minimal_tuple(), HJTUPLE_MINTUPLE, HJTUPLE_OVERHEAD, ParallelHashJoinBatchAccessor::inner_tuples, MAXALIGN, ParallelHashJoinBatchAccessor::ntuples, HashJoinTableData::parallel_state, PHJ_BUILD_HASH_INNER, ParallelHashJoinBatchAccessor::preallocated, HashJoinTableData::shared, sts_puttuple(), and MinimalTupleData::t_len.

Referenced by MultiExecParallelHash().

ExecParallelHashTableInsertCurrentBatch()

void ExecParallelHashTableInsertCurrentBatch	(	HashJoinTable	hashtable,
		TupleTableSlot *	slot,
		uint32	hashvalue
	)

Definition at line 1763 of file nodeHash.c.

{
    bool        shouldFree;
    MinimalTuple tuple = ExecFetchSlotMinimalTuple(slot, &shouldFree);
    HashJoinTuple hashTuple;
    dsa_pointer shared;
    int         batchno;
    int         bucketno;
 
    ExecHashGetBucketAndBatch(hashtable, hashvalue, &bucketno, &batchno);
    Assert(batchno == hashtable->curbatch);
    hashTuple = ExecParallelHashTupleAlloc(hashtable,
                                           HJTUPLE_OVERHEAD + tuple->t_len,
                                           &shared);
    hashTuple->hashvalue = hashvalue;
    memcpy(HJTUPLE_MINTUPLE(hashTuple), tuple, tuple->t_len);
    HeapTupleHeaderClearMatch(HJTUPLE_MINTUPLE(hashTuple));
    ExecParallelHashPushTuple(&hashtable->buckets.shared[bucketno],
                              hashTuple, shared);
 
    if (shouldFree)
        heap_free_minimal_tuple(tuple);
}

References Assert(), HashJoinTableData::buckets, HashJoinTableData::curbatch, ExecFetchSlotMinimalTuple(), ExecHashGetBucketAndBatch(), ExecParallelHashPushTuple(), ExecParallelHashTupleAlloc(), HashJoinTupleData::hashvalue, heap_free_minimal_tuple(), HeapTupleHeaderClearMatch, HJTUPLE_MINTUPLE, HJTUPLE_OVERHEAD, HashJoinTableData::shared, and MinimalTupleData::t_len.

Referenced by ExecParallelHashJoinNewBatch().

ExecParallelHashTableSetCurrentBatch()

void ExecParallelHashTableSetCurrentBatch	(	HashJoinTable	hashtable,
		int	batchno
	)

Definition at line 3284 of file nodeHash.c.

{
    Assert(hashtable->batches[batchno].shared->buckets != InvalidDsaPointer);
 
    hashtable->curbatch = batchno;
    hashtable->buckets.shared = (dsa_pointer_atomic *)
        dsa_get_address(hashtable->area,
                        hashtable->batches[batchno].shared->buckets);
    hashtable->nbuckets = hashtable->parallel_state->nbuckets;
    hashtable->log2_nbuckets = my_log2(hashtable->nbuckets);
    hashtable->current_chunk = NULL;
    hashtable->current_chunk_shared = InvalidDsaPointer;
    hashtable->batches[batchno].at_least_one_chunk = false;
}

References HashJoinTableData::area, Assert(), ParallelHashJoinBatchAccessor::at_least_one_chunk, HashJoinTableData::batches, ParallelHashJoinBatch::buckets, HashJoinTableData::buckets, HashJoinTableData::curbatch, HashJoinTableData::current_chunk, HashJoinTableData::current_chunk_shared, dsa_get_address(), InvalidDsaPointer, HashJoinTableData::log2_nbuckets, my_log2(), ParallelHashJoinState::nbuckets, HashJoinTableData::nbuckets, HashJoinTableData::parallel_state, ParallelHashJoinBatchAccessor::shared, and HashJoinTableData::shared.

Referenced by ExecParallelHashIncreaseNumBatches(), ExecParallelHashIncreaseNumBuckets(), ExecParallelHashJoinNewBatch(), and MultiExecParallelHash().

ExecParallelScanHashBucket()

bool ExecParallelScanHashBucket	(	HashJoinState *	hjstate,
		ExprContext *	econtext
	)

Definition at line 2008 of file nodeHash.c.

{
    ExprState  *hjclauses = hjstate->hashclauses;
    HashJoinTable hashtable = hjstate->hj_HashTable;
    HashJoinTuple hashTuple = hjstate->hj_CurTuple;
    uint32      hashvalue = hjstate->hj_CurHashValue;
 
    /*
     * hj_CurTuple is the address of the tuple last returned from the current
     * bucket, or NULL if it's time to start scanning a new bucket.
     */
    if (hashTuple != NULL)
        hashTuple = ExecParallelHashNextTuple(hashtable, hashTuple);
    else
        hashTuple = ExecParallelHashFirstTuple(hashtable,
                                               hjstate->hj_CurBucketNo);
 
    while (hashTuple != NULL)
    {
        if (hashTuple->hashvalue == hashvalue)
        {
            TupleTableSlot *inntuple;
 
            /* insert hashtable's tuple into exec slot so ExecQual sees it */
            inntuple = ExecStoreMinimalTuple(HJTUPLE_MINTUPLE(hashTuple),
                                             hjstate->hj_HashTupleSlot,
                                             false);    /* do not pfree */
            econtext->ecxt_innertuple = inntuple;
 
            if (ExecQualAndReset(hjclauses, econtext))
            {
                hjstate->hj_CurTuple = hashTuple;
                return true;
            }
        }
 
        hashTuple = ExecParallelHashNextTuple(hashtable, hashTuple);
    }
 
    /*
     * no match
     */
    return false;
}

References ExprContext::ecxt_innertuple, ExecParallelHashFirstTuple(), ExecParallelHashNextTuple(), ExecQualAndReset(), ExecStoreMinimalTuple(), HashJoinState::hashclauses, HashJoinTupleData::hashvalue, HashJoinState::hj_CurBucketNo, HashJoinState::hj_CurHashValue, HashJoinState::hj_CurTuple, HashJoinState::hj_HashTable, HashJoinState::hj_HashTupleSlot, and HJTUPLE_MINTUPLE.

Referenced by ExecHashJoinImpl().

ExecPrepHashTableForUnmatched()

void ExecPrepHashTableForUnmatched

(

HashJoinState *

hjstate

)

Definition at line 2059 of file nodeHash.c.

{
    /*----------
     * During this scan we use the HashJoinState fields as follows:
     *
     * hj_CurBucketNo: next regular bucket to scan
     * hj_CurSkewBucketNo: next skew bucket (an index into skewBucketNums)
     * hj_CurTuple: last tuple returned, or NULL to start next bucket
     *----------
     */
    hjstate->hj_CurBucketNo = 0;
    hjstate->hj_CurSkewBucketNo = 0;
    hjstate->hj_CurTuple = NULL;
}

References HashJoinState::hj_CurBucketNo, HashJoinState::hj_CurSkewBucketNo, and HashJoinState::hj_CurTuple.

Referenced by ExecHashJoinImpl().

ExecReScanHash()

void ExecReScanHash

(

HashState *

node

)

Definition at line 2208 of file nodeHash.c.

{
    PlanState  *outerPlan = outerPlanState(node);
 
    /*
     * if chgParam of subnode is not null then plan will be re-scanned by
     * first ExecProcNode.
     */
    if (outerPlan->chgParam == NULL)
        ExecReScan(outerPlan);
}

References ExecReScan(), outerPlan, and outerPlanState.

Referenced by ExecReScan().

ExecScanHashBucket()

bool ExecScanHashBucket	(	HashJoinState *	hjstate,
		ExprContext *	econtext
	)

Definition at line 1947 of file nodeHash.c.

{
    ExprState  *hjclauses = hjstate->hashclauses;
    HashJoinTable hashtable = hjstate->hj_HashTable;
    HashJoinTuple hashTuple = hjstate->hj_CurTuple;
    uint32      hashvalue = hjstate->hj_CurHashValue;
 
    /*
     * hj_CurTuple is the address of the tuple last returned from the current
     * bucket, or NULL if it's time to start scanning a new bucket.
     *
     * If the tuple hashed to a skew bucket then scan the skew bucket
     * otherwise scan the standard hashtable bucket.
     */
    if (hashTuple != NULL)
        hashTuple = hashTuple->next.unshared;
    else if (hjstate->hj_CurSkewBucketNo != INVALID_SKEW_BUCKET_NO)
        hashTuple = hashtable->skewBucket[hjstate->hj_CurSkewBucketNo]->tuples;
    else
        hashTuple = hashtable->buckets.unshared[hjstate->hj_CurBucketNo];
 
    while (hashTuple != NULL)
    {
        if (hashTuple->hashvalue == hashvalue)
        {
            TupleTableSlot *inntuple;
 
            /* insert hashtable's tuple into exec slot so ExecQual sees it */
            inntuple = ExecStoreMinimalTuple(HJTUPLE_MINTUPLE(hashTuple),
                                             hjstate->hj_HashTupleSlot,
                                             false);    /* do not pfree */
            econtext->ecxt_innertuple = inntuple;
 
            if (ExecQualAndReset(hjclauses, econtext))
            {
                hjstate->hj_CurTuple = hashTuple;
                return true;
            }
        }
 
        hashTuple = hashTuple->next.unshared;
    }
 
    /*
     * no match
     */
    return false;
}

References HashJoinTableData::buckets, ExprContext::ecxt_innertuple, ExecQualAndReset(), ExecStoreMinimalTuple(), HashJoinState::hashclauses, HashJoinTupleData::hashvalue, HashJoinState::hj_CurBucketNo, HashJoinState::hj_CurHashValue, HashJoinState::hj_CurSkewBucketNo, HashJoinState::hj_CurTuple, HashJoinState::hj_HashTable, HashJoinState::hj_HashTupleSlot, HJTUPLE_MINTUPLE, INVALID_SKEW_BUCKET_NO, HashJoinTupleData::next, HashJoinTableData::skewBucket, HashSkewBucket::tuples, HashJoinTupleData::unshared, and HashJoinTableData::unshared.

Referenced by ExecHashJoinImpl().

ExecScanHashTableForUnmatched()

bool ExecScanHashTableForUnmatched	(	HashJoinState *	hjstate,
		ExprContext *	econtext
	)

Definition at line 2083 of file nodeHash.c.

{
    HashJoinTable hashtable = hjstate->hj_HashTable;
    HashJoinTuple hashTuple = hjstate->hj_CurTuple;
 
    for (;;)
    {
        /*
         * hj_CurTuple is the address of the tuple last returned from the
         * current bucket, or NULL if it's time to start scanning a new
         * bucket.
         */
        if (hashTuple != NULL)
            hashTuple = hashTuple->next.unshared;
        else if (hjstate->hj_CurBucketNo < hashtable->nbuckets)
        {
            hashTuple = hashtable->buckets.unshared[hjstate->hj_CurBucketNo];
            hjstate->hj_CurBucketNo++;
        }
        else if (hjstate->hj_CurSkewBucketNo < hashtable->nSkewBuckets)
        {
            int         j = hashtable->skewBucketNums[hjstate->hj_CurSkewBucketNo];
 
            hashTuple = hashtable->skewBucket[j]->tuples;
            hjstate->hj_CurSkewBucketNo++;
        }
        else
            break;              /* finished all buckets */
 
        while (hashTuple != NULL)
        {
            if (!HeapTupleHeaderHasMatch(HJTUPLE_MINTUPLE(hashTuple)))
            {
                TupleTableSlot *inntuple;
 
                /* insert hashtable's tuple into exec slot */
                inntuple = ExecStoreMinimalTuple(HJTUPLE_MINTUPLE(hashTuple),
                                                 hjstate->hj_HashTupleSlot,
                                                 false);    /* do not pfree */
                econtext->ecxt_innertuple = inntuple;
 
                /*
                 * Reset temp memory each time; although this function doesn't
                 * do any qual eval, the caller will, so let's keep it
                 * parallel to ExecScanHashBucket.
                 */
                ResetExprContext(econtext);
 
                hjstate->hj_CurTuple = hashTuple;
                return true;
            }
 
            hashTuple = hashTuple->next.unshared;
        }
 
        /* allow this loop to be cancellable */
        CHECK_FOR_INTERRUPTS();
    }
 
    /*
     * no more unmatched tuples
     */
    return false;
}

References HashJoinTableData::buckets, CHECK_FOR_INTERRUPTS, ExprContext::ecxt_innertuple, ExecStoreMinimalTuple(), HeapTupleHeaderHasMatch, HashJoinState::hj_CurBucketNo, HashJoinState::hj_CurSkewBucketNo, HashJoinState::hj_CurTuple, HashJoinState::hj_HashTable, HashJoinState::hj_HashTupleSlot, HJTUPLE_MINTUPLE, j, HashJoinTableData::nbuckets, HashJoinTupleData::next, HashJoinTableData::nSkewBuckets, ResetExprContext, HashJoinTableData::skewBucket, HashJoinTableData::skewBucketNums, HashSkewBucket::tuples, HashJoinTupleData::unshared, and HashJoinTableData::unshared.

Referenced by ExecHashJoinImpl().

ExecShutdownHash()

void ExecShutdownHash

(

HashState *

node

)

Definition at line 2658 of file nodeHash.c.

{
    /* Allocate save space if EXPLAIN'ing and we didn't do so already */
    if (node->ps.instrument && !node->hinstrument)
        node->hinstrument = palloc0_object(HashInstrumentation);
    /* Now accumulate data for the current (final) hash table */
    if (node->hinstrument && node->hashtable)
        ExecHashAccumInstrumentation(node->hinstrument, node->hashtable);
}

References ExecHashAccumInstrumentation(), HashState::hashtable, HashState::hinstrument, PlanState::instrument, palloc0_object, and HashState::ps.

Referenced by ExecShutdownNode_walker().

MultiExecHash()

Node* MultiExecHash

(

HashState *

node

)

Definition at line 106 of file nodeHash.c.

{
    /* must provide our own instrumentation support */
    if (node->ps.instrument)
        InstrStartNode(node->ps.instrument);
 
    if (node->parallel_state != NULL)
        MultiExecParallelHash(node);
    else
        MultiExecPrivateHash(node);
 
    /* must provide our own instrumentation support */
    if (node->ps.instrument)
        InstrStopNode(node->ps.instrument, node->hashtable->partialTuples);
 
    /*
     * We do not return the hash table directly because it's not a subtype of
     * Node, and so would violate the MultiExecProcNode API.  Instead, our
     * parent Hashjoin node is expected to know how to fish it out of our node
     * state.  Ugly but not really worth cleaning up, since Hashjoin knows
     * quite a bit more about Hash besides that.
     */
    return NULL;
}

References HashState::hashtable, InstrStartNode(), InstrStopNode(), PlanState::instrument, MultiExecParallelHash(), MultiExecPrivateHash(), HashState::parallel_state, HashJoinTableData::partialTuples, and HashState::ps.

Referenced by MultiExecProcNode().