[BugFix] fix cte output column duplicate (#34904)

Signed-off-by: Seaven <seaven_7@qq.com>
StarRocks · Nov 15, 2023 · a2015c1 · a2015c1
1 parent 1bbb88f
commit a2015c1
Show file tree

Hide file tree

Showing 3 changed files with 56 additions and 2 deletions.
diff --git a/fe/fe-core/src/main/java/com/starrocks/sql/plan/PlanFragmentBuilder.java b/fe/fe-core/src/main/java/com/starrocks/sql/plan/PlanFragmentBuilder.java
@@ -739,7 +739,8 @@ public PlanFragment visitPhysicalOlapScan(OptExpression optExpr, ExecPlan contex
 
                     for (PhysicalPartition physicalPartition : partition.getSubPartitions()) {
                         Map<Long, Integer> tabletId2BucketSeq = Maps.newHashMap();
-                        List<Long> selectTabletIds = scanNode.getPartitionToScanTabletMap().get(physicalPartition.getId());
+                        List<Long> selectTabletIds = scanNode.getPartitionToScanTabletMap()
+                                .get(physicalPartition.getId());
                         Preconditions.checkState(selectTabletIds != null && !selectTabletIds.isEmpty());
                         final MaterializedIndex selectedTable = physicalPartition.getIndex(selectedIndexId);
                         List<Long> allTabletIds = selectedTable.getTabletIdsInOrder();
@@ -2796,7 +2797,7 @@ public PlanFragment visitPhysicalCTEConsume(OptExpression optExpression, ExecPla
                     cteFragment.getPlanRoot(), DistributionSpec.DistributionType.SHUFFLE);
 
             exchangeNode.setReceiveColumns(consume.getCteOutputColumnRefMap().values().stream()
-                    .map(ColumnRefOperator::getId).collect(Collectors.toList()));
+                    .map(ColumnRefOperator::getId).distinct().collect(Collectors.toList()));
             exchangeNode.setDataPartition(cteFragment.getDataPartition());
 
             exchangeNode.setNumInstances(cteFragment.getPlanRoot().getNumInstances());

diff --git a/test/sql/test_subquery/R/test_complex_in b/test/sql/test_subquery/R/test_complex_in
@@ -0,0 +1,29 @@
+-- name: test_same_correlation_col_in
+CREATE TABLE `t0` (
+  `v1` bigint(20) NULL COMMENT "",
+  `v2` bigint(20) NULL COMMENT "",
+  `v3` bigint(20) NULL COMMENT "",
+  `v4` largeint NULL COMMENT "",
+  `v5` tinyint NULL COMMENT ""
+) ENGINE=OLAP
+DUPLICATE KEY(`v1`)
+COMMENT "OLAP"
+DISTRIBUTED BY HASH(`v1`) BUCKETS 3
+PROPERTIES (
+"replication_num" = "1",
+"in_memory" = "false",
+"enable_persistent_index" = "false",
+"replicated_storage" = "false",
+"compression" = "LZ4"
+);
+-- result:
+-- !result
+insert into t0 values (1, 1, 1, null, null), (2, 2, 2, null, null);
+-- result:
+-- !result
+select * from t0 
+where (t0.v2 = 2) or t0.v3 in (select x0.v3 from t0 x0 where t0.v2 = x0.v3);
+-- result:
+2	2	2	None	None
+1	1	1	None	None
+-- !result
diff --git a/test/sql/test_subquery/T/test_complex_in b/test/sql/test_subquery/T/test_complex_in
@@ -0,0 +1,24 @@
+-- name: test_same_correlation_col_in
+CREATE TABLE `t0` (
+  `v1` bigint(20) NULL COMMENT "",
+  `v2` bigint(20) NULL COMMENT "",
+  `v3` bigint(20) NULL COMMENT "",
+  `v4` largeint NULL COMMENT "",
+  `v5` tinyint NULL COMMENT ""
+) ENGINE=OLAP
+DUPLICATE KEY(`v1`)
+COMMENT "OLAP"
+DISTRIBUTED BY HASH(`v1`) BUCKETS 3
+PROPERTIES (
+"replication_num" = "1",
+"in_memory" = "false",
+"enable_persistent_index" = "false",
+"replicated_storage" = "false",
+"compression" = "LZ4"
+);
+
+insert into t0 values (1, 1, 1, null, null), (2, 2, 2, null, null);
+
+
+select * from t0 
+where (t0.v2 = 2) or t0.v3 in (select x0.v3 from t0 x0 where t0.v2 = x0.v3);