Remove 'additional' pointer from TupleHashEntryData.

jeff-davis · jeff-davis · commit 626df47ad9db · 2025-03-24T22:06:02.000-07:00
Reduces memory required for hash aggregation by avoiding an allocation and a pointer in the TupleHashEntryData structure. That structure is used for all buckets, whether occupied or not, so the savings is substantial. Discussion: https://postgr.es/m/AApHDvpN4v3t_sdz4dvrv1Fx_ZPw=twSnxuTEytRYP7LFz5K9A@mail.gmail.com Reviewed-by: David Rowley <dgrowleyml@gmail.com>
diff --git a/src/backend/executor/execGrouping.c b/src/backend/executor/execGrouping.c
@@ -485,11 +485,18 @@ LookupTupleHashEntry_internal(TupleHashTable hashtable, TupleTableSlot *slot,
 
 			MemoryContextSwitchTo(hashtable->tablecxt);
 
-			entry->firstTuple = ExecCopySlotMinimalTuple(slot);
-			if (hashtable->additionalsize > 0)
-				entry->additional = palloc0(hashtable->additionalsize);
-			else
-				entry->additional = NULL;
+			/*
+			 * Copy the first tuple into the table context, and request
+			 * additionalsize extra bytes before the allocation.
+			 *
+			 * The caller can get a pointer to the additional data with
+			 * TupleHashEntryGetAdditional(), and store arbitrary data there.
+			 * Placing both the tuple and additional data in the same
+			 * allocation avoids the need to store an extra pointer in
+			 * TupleHashEntryData or allocate an additional chunk.
+			 */
+			entry->firstTuple = ExecCopySlotMinimalTupleExtra(slot,
+															  hashtable->additionalsize);
 		}
 	}
 	else
diff --git a/src/include/executor/executor.h b/src/include/executor/executor.h
@@ -188,7 +188,10 @@ TupleHashEntryGetTuple(TupleHashEntry entry)
 static inline void *
 TupleHashEntryGetAdditional(TupleHashTable hashtable, TupleHashEntry entry)
 {
-	return entry->additional;
+	if (hashtable->additionalsize > 0)
+		return (char *) entry->firstTuple - hashtable->additionalsize;
+	else
+		return NULL;
 }
 #endif
 
diff --git a/src/include/nodes/execnodes.h b/src/include/nodes/execnodes.h
@@ -840,7 +840,6 @@ typedef struct TupleHashTableData *TupleHashTable;
 typedef struct TupleHashEntryData
 {
 	MinimalTuple firstTuple;	/* copy of first tuple in this group */
-	void	   *additional;		/* user data */
 	uint32		status;			/* hash status */
 	uint32		hash;			/* hash value (cached) */
 } TupleHashEntryData;

Original file line number	Diff line number	Diff line change
`@@ -188,7 +188,10 @@ TupleHashEntryGetTuple(TupleHashEntry entry)`
`188`	`188`	`static inline void *`
`189`	`189`	`TupleHashEntryGetAdditional(TupleHashTable hashtable, TupleHashEntry entry)`
`190`	`190`	`{`
`191`		`- return entry->additional;`
	`191`	`+ if (hashtable->additionalsize > 0)`
	`192`	`+ return (char *) entry->firstTuple - hashtable->additionalsize;`
	`193`	`+ else`
	`194`	`+ return NULL;`
`192`	`195`	`}`
`193`	`196`	`#endif`
`194`	`197`
Original file line number	Diff line number	Diff line change
`@@ -840,7 +840,6 @@ typedef struct TupleHashTableData *TupleHashTable;`
`840`	`840`	`typedef struct TupleHashEntryData`
`841`	`841`	`{`
`842`	`842`	`MinimalTuple firstTuple; /* copy of first tuple in this group */`
`843`		`- void additional; / user data */`
`844`	`843`	`uint32 status; /* hash status */`
`845`	`844`	`uint32 hash; /* hash value (cached) */`
`846`	`845`	`} TupleHashEntryData;`