Use column collation for extended statistics

tvondra · tvondra · commit a63378a03ec0 · 2019-07-20T16:37:37.000+02:00
The current extended statistics code was a bit confused which collation to use. When building the statistics, the collations defined as default for the data types were used (since commit 5e09280). The MCV code was however using the column collations for MCV serialization, and then DEFAULT_COLLATION_OID when computing estimates. So overall the code was using all three possible options, inconsistently. This uses the column colation everywhere - this makes it consistent with what 5e09280 did for regular stats. We however do not track the collations in a catalog, because we can derive them from column-level information. This may need to change in the future, e.g. after allowing statistics on expressions. Reviewed-by: Tom Lane Discussion: https://postgr.es/m/8736jdhbhc.fsf%40ansel.ydns.eu Backpatch-to: 12
diff --git a/src/backend/commands/statscmds.c b/src/backend/commands/statscmds.c
@@ -485,6 +485,10 @@ RemoveStatisticsById(Oid statsOid)
  *
  * For MCV lists that's not the case, as those statistics store the datums
  * internally. In this case we simply reset the statistics value to NULL.
+ *
+ * Note that "type change" includes collation change, which means we can rely
+ * on the MCV list being consistent with the collation info in pg_attribute
+ * during estimation.
  */
 void
 UpdateStatisticsForTypeChange(Oid statsOid, Oid relationOid, int attnum,
diff --git a/src/backend/statistics/dependencies.c b/src/backend/statistics/dependencies.c
@@ -273,7 +273,7 @@ dependency_degree(int numrows, HeapTuple *rows, int k, AttrNumber *dependency,
 				 colstat->attrtypid);
 
 		/* prepare the sort function for this dimension */
-		multi_sort_add_dimension(mss, i, type->lt_opr, type->typcollation);
+		multi_sort_add_dimension(mss, i, type->lt_opr, colstat->attrcollid);
 	}
 
 	/*
diff --git a/src/backend/statistics/mcv.c b/src/backend/statistics/mcv.c
@@ -366,7 +366,7 @@ build_mss(VacAttrStats **stats, int numattrs)
 			elog(ERROR, "cache lookup failed for ordering operator for type %u",
 				 colstat->attrtypid);
 
-		multi_sort_add_dimension(mss, i, type->lt_opr, type->typcollation);
+		multi_sort_add_dimension(mss, i, type->lt_opr, colstat->attrcollid);
 	}
 
 	return mss;
@@ -686,7 +686,7 @@ statext_mcv_serialize(MCVList *mcvlist, VacAttrStats **stats)
 
 		/* sort and deduplicate the data */
 		ssup[dim].ssup_cxt = CurrentMemoryContext;
-		ssup[dim].ssup_collation = DEFAULT_COLLATION_OID;
+		ssup[dim].ssup_collation = stats[dim]->attrcollid;
 		ssup[dim].ssup_nulls_first = false;
 
 		PrepareSortSupportFromOrderingOp(typentry->lt_opr, &ssup[dim]);
@@ -1630,15 +1630,22 @@ mcv_get_match_bitmap(PlannerInfo *root, List *clauses,
 					 * First check whether the constant is below the lower
 					 * boundary (in that case we can skip the bucket, because
 					 * there's no overlap).
+					 *
+					 * We don't store collations used to build the statistics,
+					 * but we can use the collation for the attribute itself,
+					 * as stored in varcollid. We do reset the statistics after
+					 * a type change (including collation change), so this is
+					 * OK. We may need to relax this after allowing extended
+					 * statistics on expressions.
 					 */
 					if (varonleft)
 						match = DatumGetBool(FunctionCall2Coll(&opproc,
-															   DEFAULT_COLLATION_OID,
+															   var->varcollid,
 															   item->values[idx],
 															   cst->constvalue));
 					else
 						match = DatumGetBool(FunctionCall2Coll(&opproc,
-															   DEFAULT_COLLATION_OID,
+															   var->varcollid,
 															   cst->constvalue,
 															   item->values[idx]));
 
diff --git a/src/backend/statistics/mvdistinct.c b/src/backend/statistics/mvdistinct.c
@@ -477,7 +477,7 @@ ndistinct_for_combination(double totalrows, int numrows, HeapTuple *rows,
 				 colstat->attrtypid);
 
 		/* prepare the sort function for this dimension */
-		multi_sort_add_dimension(mss, i, type->lt_opr, type->typcollation);
+		multi_sort_add_dimension(mss, i, type->lt_opr, colstat->attrcollid);
 
 		/* accumulate all the data for this dimension into the arrays */
 		for (j = 0; j < numrows; j++)

Original file line number	Diff line number	Diff line change
`@@ -273,7 +273,7 @@ dependency_degree(int numrows, HeapTuple rows, int k, AttrNumber dependency,`
`273`	`273`	`colstat->attrtypid);`
`274`	`274`
`275`	`275`	`/* prepare the sort function for this dimension */`
`276`		`- multi_sort_add_dimension(mss, i, type->lt_opr, type->typcollation);`
	`276`	`+ multi_sort_add_dimension(mss, i, type->lt_opr, colstat->attrcollid);`
`277`	`277`	`}`
`278`	`278`
`279`	`279`	`/*`