Create Block Directory for Materialized Views of AO storage.

Fix: apache#865

Implemented a solution that creates a block directory for materialized
views using AO/AOCS storage with an index.
This fix prevents errors that occur when refreshing materialized views
in AO/AOCS storage.

Authored-by: Zhang Mingli [email protected]

Author: avamingli

src/backend/commands/matview.c

@@ -410,6 +410,7 @@ ExecRefreshMatView(RefreshMatViewStmt *stmt, const char *queryString,
 	ObjectAddress address;
 	RefreshClause *refreshClause;
 	bool oldPopulated;
+	bool ao_has_index;
 
 	/* MATERIALIZED_VIEW_FIXME: Refresh MatView is not MPP-fied. */
 
@@ -649,13 +650,19 @@ ExecRefreshMatView(RefreshMatViewStmt *stmt, const char *queryString,
 		free_object_addresses(immv_triggers);
 	}
 
+	/*
+	 * Fix issue https://github.com/apache/cloudberry/issues/865
+	 * Create blockdir for Matartalized View of AO/AOCS storage has index.
+	 */
+	ao_has_index = matviewRel->rd_rel->relhasindex && RelationIsAppendOptimized(matviewRel);
+
 	/*
 	 * Create the transient table that will receive the regenerated data. Lock
 	 * it against access by any other process until commit (by which time it
 	 * will be gone).
 	 */
 	OIDNewHeap = make_new_heap(matviewOid, tableSpace, matviewRel->rd_rel->relam, relpersistence,
-							   ExclusiveLock, false, true);
+							   ExclusiveLock, ao_has_index, true);
 	LockRelationOid(OIDNewHeap, AccessExclusiveLock);
 	dest = CreateTransientRelDestReceiver(OIDNewHeap, matviewOid, concurrent, relpersistence,
 										  stmt->skipData);
@@ -888,6 +895,7 @@ transientrel_init(QueryDesc *queryDesc)
 	char		relpersistence;
 	LOCKMODE	lockmode;
 	RefreshClause *refreshClause;
+	bool		ao_has_index;
 
 	refreshClause = queryDesc->plannedstmt->refreshClause;
 	/* Determine strength of lock needed. */
@@ -919,14 +927,17 @@ transientrel_init(QueryDesc *queryDesc)
 		tableSpace = matviewRel->rd_rel->reltablespace;
 		relpersistence = matviewRel->rd_rel->relpersistence;
 	}
+
+	ao_has_index = matviewRel->rd_rel->relhasindex && RelationIsAppendOptimized(matviewRel);
+
 	/*
 	 * Create the transient table that will receive the regenerated data. Lock
 	 * it against access by any other process until commit (by which time it
 	 * will be gone).
 	 */
 	OIDNewHeap = make_new_heap(matviewOid, tableSpace, matviewRel->rd_rel->relam,
 							   relpersistence,
-							   ExclusiveLock, false, false);
+							   ExclusiveLock, ao_has_index, false);
 	LockRelationOid(OIDNewHeap, AccessExclusiveLock);
 
 	queryDesc->dest = CreateTransientRelDestReceiver(OIDNewHeap, matviewOid, concurrent,

src/test/regress/expected/matview.out

@@ -809,3 +809,50 @@ ERROR:  materialized view "mat_view_twn" has not been populated
 HINT:  Use the REFRESH MATERIALIZED VIEW command.
 DROP MATERIALIZED VIEW mat_view_twn;
 DROP TABLE mvtest_twn;
+--
+-- https://github.com/apache/cloudberry/issues/865
+--
+set default_table_access_method TO AO_ROW;
+CREATE TABLE t_issue_865_ao
+(
+    id           bigint NOT NULL,
+    user_id      bigint
+);
+NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause -- Using column named 'id' as the Cloudberry Database data distribution key for this table.
+HINT:  The 'DISTRIBUTED BY' clause determines the distribution of data. Make sure column(s) chosen are the optimal data distribution key to minimize skew.
+insert into t_issue_865_ao values (1, 1), (2, 1), (3, 2), (4, 2), (5, 3), (6, 3), (7, 4), (8, 4), (9, 5), (10, 5);
+CREATE MATERIALIZED VIEW matview_issue_865_ao AS SELECT * FROM t_issue_865_ao WHERE id < 6;
+NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause. Creating a NULL policy entry.
+CREATE INDEX idx_matview_issue_865_ao ON matview_issue_865_ao USING btree (user_id);
+BEGIN;
+UPDATE t_issue_865_ao SET id = id WHERE id = 1;
+UPDATE t_issue_865_ao SET id = id WHERE id = 2;
+UPDATE t_issue_865_ao SET id = id WHERE id = 3;
+COMMIT;
+VACUUM t_issue_865_ao;
+REFRESH MATERIALIZED VIEW matview_issue_865_ao;
+-- AOCS
+set default_table_access_method TO AO_COLUMN;
+CREATE TABLE t_issue_865_aocs
+(
+    id           bigint NOT NULL,
+    user_id      bigint
+);
+NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause -- Using column named 'id' as the Cloudberry Database data distribution key for this table.
+HINT:  The 'DISTRIBUTED BY' clause determines the distribution of data. Make sure column(s) chosen are the optimal data distribution key to minimize skew.
+insert into t_issue_865_aocs values (1, 1), (2, 1), (3, 2), (4, 2), (5, 3), (6, 3), (7, 4), (8, 4), (9, 5), (10, 5);
+CREATE MATERIALIZED VIEW matview_issue_865_aocs AS SELECT * FROM t_issue_865_aocs WHERE id < 6;
+NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause. Creating a NULL policy entry.
+CREATE INDEX idx_matview_issue_865_aocs ON matview_issue_865_aocs USING btree (user_id);
+BEGIN;
+UPDATE t_issue_865_aocs SET id = id WHERE id = 1;
+UPDATE t_issue_865_aocs SET id = id WHERE id = 2;
+UPDATE t_issue_865_aocs SET id = id WHERE id = 3;
+COMMIT;
+VACUUM t_issue_865_aocs;
+REFRESH MATERIALIZED VIEW matview_issue_865_aocs;
+RESET default_table_access_method;
+DROP TABLE t_issue_865_ao CASCADE; 
+NOTICE:  drop cascades to materialized view matview_issue_865_ao
+DROP TABLE t_issue_865_aocs CASCADE; 
+NOTICE:  drop cascades to materialized view matview_issue_865_aocs

src/test/regress/expected/matview_optimizer.out

@@ -826,3 +826,50 @@ ERROR:  materialized view "mat_view_twn" has not been populated
 HINT:  Use the REFRESH MATERIALIZED VIEW command.
 DROP MATERIALIZED VIEW mat_view_twn;
 DROP TABLE mvtest_twn;
+--
+-- https://github.com/apache/cloudberry/issues/865
+--
+set default_table_access_method TO AO_ROW;
+CREATE TABLE t_issue_865_ao
+(
+    id           bigint NOT NULL,
+    user_id      bigint
+);
+NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause -- Using column named 'id' as the Cloudberry Database data distribution key for this table.
+HINT:  The 'DISTRIBUTED BY' clause determines the distribution of data. Make sure column(s) chosen are the optimal data distribution key to minimize skew.
+insert into t_issue_865_ao values (1, 1), (2, 1), (3, 2), (4, 2), (5, 3), (6, 3), (7, 4), (8, 4), (9, 5), (10, 5);
+CREATE MATERIALIZED VIEW matview_issue_865_ao AS SELECT * FROM t_issue_865_ao WHERE id < 6;
+NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause. Creating a NULL policy entry.
+CREATE INDEX idx_matview_issue_865_ao ON matview_issue_865_ao USING btree (user_id);
+BEGIN;
+UPDATE t_issue_865_ao SET id = id WHERE id = 1;
+UPDATE t_issue_865_ao SET id = id WHERE id = 2;
+UPDATE t_issue_865_ao SET id = id WHERE id = 3;
+COMMIT;
+VACUUM t_issue_865_ao;
+REFRESH MATERIALIZED VIEW matview_issue_865_ao;
+-- AOCS
+set default_table_access_method TO AO_COLUMN;
+CREATE TABLE t_issue_865_aocs
+(
+    id           bigint NOT NULL,
+    user_id      bigint
+);
+NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause -- Using column named 'id' as the Cloudberry Database data distribution key for this table.
+HINT:  The 'DISTRIBUTED BY' clause determines the distribution of data. Make sure column(s) chosen are the optimal data distribution key to minimize skew.
+insert into t_issue_865_aocs values (1, 1), (2, 1), (3, 2), (4, 2), (5, 3), (6, 3), (7, 4), (8, 4), (9, 5), (10, 5);
+CREATE MATERIALIZED VIEW matview_issue_865_aocs AS SELECT * FROM t_issue_865_aocs WHERE id < 6;
+NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause. Creating a NULL policy entry.
+CREATE INDEX idx_matview_issue_865_aocs ON matview_issue_865_aocs USING btree (user_id);
+BEGIN;
+UPDATE t_issue_865_aocs SET id = id WHERE id = 1;
+UPDATE t_issue_865_aocs SET id = id WHERE id = 2;
+UPDATE t_issue_865_aocs SET id = id WHERE id = 3;
+COMMIT;
+VACUUM t_issue_865_aocs;
+REFRESH MATERIALIZED VIEW matview_issue_865_aocs;
+RESET default_table_access_method;
+DROP TABLE t_issue_865_ao CASCADE; 
+NOTICE:  drop cascades to materialized view matview_issue_865_ao
+DROP TABLE t_issue_865_aocs CASCADE; 
+NOTICE:  drop cascades to materialized view matview_issue_865_aocs

src/test/regress/sql/matview.sql

@@ -358,3 +358,55 @@ SELECT * FROM mat_view_twn;
 
 DROP MATERIALIZED VIEW mat_view_twn;
 DROP TABLE mvtest_twn;
+
+--
+-- https://github.com/apache/cloudberry/issues/865
+--
+set default_table_access_method TO AO_ROW;
+
+CREATE TABLE t_issue_865_ao
+(
+    id           bigint NOT NULL,
+    user_id      bigint
+);
+insert into t_issue_865_ao values (1, 1), (2, 1), (3, 2), (4, 2), (5, 3), (6, 3), (7, 4), (8, 4), (9, 5), (10, 5);
+
+CREATE MATERIALIZED VIEW matview_issue_865_ao AS SELECT * FROM t_issue_865_ao WHERE id < 6;
+CREATE INDEX idx_matview_issue_865_ao ON matview_issue_865_ao USING btree (user_id);
+
+BEGIN;
+UPDATE t_issue_865_ao SET id = id WHERE id = 1;
+UPDATE t_issue_865_ao SET id = id WHERE id = 2;
+UPDATE t_issue_865_ao SET id = id WHERE id = 3;
+COMMIT;
+
+VACUUM t_issue_865_ao;
+
+REFRESH MATERIALIZED VIEW matview_issue_865_ao;
+
+-- AOCS
+set default_table_access_method TO AO_COLUMN;
+
+CREATE TABLE t_issue_865_aocs
+(
+    id           bigint NOT NULL,
+    user_id      bigint
+);
+insert into t_issue_865_aocs values (1, 1), (2, 1), (3, 2), (4, 2), (5, 3), (6, 3), (7, 4), (8, 4), (9, 5), (10, 5);
+
+CREATE MATERIALIZED VIEW matview_issue_865_aocs AS SELECT * FROM t_issue_865_aocs WHERE id < 6;
+CREATE INDEX idx_matview_issue_865_aocs ON matview_issue_865_aocs USING btree (user_id);
+
+BEGIN;
+UPDATE t_issue_865_aocs SET id = id WHERE id = 1;
+UPDATE t_issue_865_aocs SET id = id WHERE id = 2;
+UPDATE t_issue_865_aocs SET id = id WHERE id = 3;
+COMMIT;
+
+VACUUM t_issue_865_aocs;
+
+REFRESH MATERIALIZED VIEW matview_issue_865_aocs;
+
+RESET default_table_access_method;
+DROP TABLE t_issue_865_ao CASCADE; 
+DROP TABLE t_issue_865_aocs CASCADE;