[BugFix] Fix wrong result when query cache work with select node

The query cache does not normalize the select node during the normalize stage, which can lead to incorrect results if different predicates use the same digest. Signed-off-by: stdpain <[email protected]>
stdpain · Jan 21, 2025 · 9fd94b5 · 9fd94b5
1 parent 8216577
commit 9fd94b5
Show file tree

Hide file tree

Showing 4 changed files with 74 additions and 0 deletions.
diff --git a/fe/fe-core/src/main/java/com/starrocks/planner/SelectNode.java b/fe/fe-core/src/main/java/com/starrocks/planner/SelectNode.java
@@ -37,14 +37,18 @@
 import com.starrocks.analysis.Analyzer;
 import com.starrocks.analysis.Expr;
 import com.starrocks.analysis.SlotId;
+import com.starrocks.common.Pair;
 import com.starrocks.common.StarRocksException;
 import com.starrocks.thrift.TExplainLevel;
+import com.starrocks.thrift.TNormalPlanNode;
+import com.starrocks.thrift.TNormalSelectNode;
 import com.starrocks.thrift.TPlanNode;
 import com.starrocks.thrift.TPlanNodeType;
 import com.starrocks.thrift.TSelectNode;
 import org.apache.logging.log4j.LogManager;
 import org.apache.logging.log4j.Logger;
 
+import java.nio.ByteBuffer;
 import java.util.Arrays;
 import java.util.List;
 import java.util.Map;
@@ -84,6 +88,16 @@ public void init(Analyzer analyzer) throws StarRocksException {
     public void computeStats(Analyzer analyzer) {
     }
 
+    @Override
+    protected void toNormalForm(TNormalPlanNode planNode, FragmentNormalizer normalizer) {
+        TNormalSelectNode selectNode = new TNormalSelectNode();
+        Pair<List<Integer>, List<ByteBuffer>> slotIdsAndExprs = normalizer.normalizeSlotIdsAndExprs(commonSlotMap);
+        selectNode.setCse_slot_ids(slotIdsAndExprs.first);
+        selectNode.setCse_exprs(slotIdsAndExprs.second);
+        planNode.setNode_type(TPlanNodeType.SELECT_NODE);
+        normalizeConjuncts(normalizer, planNode, conjuncts);
+    }
+
     @Override
     protected String getNodeExplainString(String prefix, TExplainLevel detailLevel) {
         StringBuilder output = new StringBuilder();

diff --git a/gensrc/thrift/Normalization.thrift b/gensrc/thrift/Normalization.thrift
@@ -143,6 +143,11 @@ struct TNormalSetOperationNode {
   4: optional i64 first_materialized_child_idx
 }
 
+struct TNormalSelectNode {
+  1: optional list<Types.TSlotId> cse_slot_ids;
+  2: optional list<binary> cse_exprs;
+}
+
 struct TNormalPlanNode {
   1: optional Types.TPlanNodeId node_id
   2: optional PlanNodes.TPlanNodeType node_type
@@ -166,4 +171,5 @@ struct TNormalPlanNode {
   19: optional TNormalSortNode sort_node
   20: optional TNormalSortAggregationNode sort_aggregation_node
   22: optional TNormalSetOperationNode set_operation_node
+  23: optional TNormalSelectNode select_node
 }
diff --git a/test/sql/test_query_cache/R/test_query_cache_select_node b/test/sql/test_query_cache/R/test_query_cache_select_node
@@ -0,0 +1,32 @@
+-- name: test_query_cache_select_node
+CREATE TABLE `tarray` (
+  `id` int(4) NULL COMMENT "",
+  `val` array<int> NULL COMMENT ""
+) ENGINE=OLAP 
+DUPLICATE KEY(`id`)
+COMMENT "OLAP"
+DISTRIBUTED BY HASH(`id`) BUCKETS 2
+PROPERTIES (
+"compression" = "LZ4",
+"fast_schema_evolution" = "false",
+"replicated_storage" = "true",
+"replication_num" = "1"
+);
+-- result:
+-- !result
+insert into tarray SELECT generate_series, [generate_series % 4, generate_series%3] FROM TABLE(generate_series(1,  4096));
+-- result:
+-- !result
+set enable_query_cache = true;
+-- result:
+-- !result
+set pipeline_dop=1;
+-- result:
+-- !result
+SELECT rule_id,COUNT(*) AS cnt FROM ( SELECT  id,unnest AS rule_id FROM tarray, unnest(val) ) er WHERE rule_id IN (28,128,127) GROUP BY  rule_id;
+-- result:
+-- !result
+SELECT rule_id,COUNT(*) AS cnt FROM ( SELECT  id,unnest AS rule_id FROM tarray, unnest(val) ) er WHERE rule_id IN (1) GROUP BY  rule_id;
+-- result:
+1	2390
+-- !result
diff --git a/test/sql/test_query_cache/T/test_query_cache_select_node b/test/sql/test_query_cache/T/test_query_cache_select_node
@@ -0,0 +1,22 @@
+-- name: test_query_cache_select_node
+CREATE TABLE `tarray` (
+  `id` int(4) NULL COMMENT "",
+  `val` array<int> NULL COMMENT ""
+) ENGINE=OLAP 
+DUPLICATE KEY(`id`)
+COMMENT "OLAP"
+DISTRIBUTED BY HASH(`id`) BUCKETS 2
+PROPERTIES (
+"compression" = "LZ4",
+"fast_schema_evolution" = "false",
+"replicated_storage" = "true",
+"replication_num" = "1"
+);
+
+insert into tarray SELECT generate_series, [generate_series % 4, generate_series%3] FROM TABLE(generate_series(1,  4096));
+
+set enable_query_cache = true;
+set pipeline_dop=1;
+
+SELECT rule_id,COUNT(*) AS cnt FROM ( SELECT  id,unnest AS rule_id FROM tarray, unnest(val) ) er WHERE rule_id IN (28,128,127) GROUP BY  rule_id;
+SELECT rule_id,COUNT(*) AS cnt FROM ( SELECT  id,unnest AS rule_id FROM tarray, unnest(val) ) er WHERE rule_id IN (1) GROUP BY  rule_id;