[GLUTEN-6669][CH] Fix diff of cast string to boolean (#6711)

apache · Aug 6, 2024 · 6ab62fc · 6ab62fc
1 parent cc1f495
commit 6ab62fc
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 1 deletion.
diff --git a/...est/scala/org/apache/gluten/execution/tpch/GlutenClickHouseTPCHSaltNullParquetSuite.scala b/...est/scala/org/apache/gluten/execution/tpch/GlutenClickHouseTPCHSaltNullParquetSuite.scala
@@ -2435,6 +2435,16 @@ class GlutenClickHouseTPCHSaltNullParquetSuite extends GlutenClickHouseTPCHAbstr
     }
   }
 
+  test("GLUTEN-6669: test cast string to boolean") {
+    withSQLConf(
+      SQLConf.OPTIMIZER_EXCLUDED_RULES.key -> (ConstantFolding.ruleName + "," + NullPropagation.ruleName)) {
+      runQueryAndCompare(
+        "select cast('1' as boolean), cast('t' as boolean), cast('all' as boolean), cast('f' as boolean)",
+        noFallBack = false
+      )(checkGlutenOperatorMatch[ProjectExecTransformer])
+    }
+  }
+
   test("GLUTEN-4032: fix shuffle read coredump after union") {
     val sql =
       """

diff --git a/cpp-ch/local-engine/Parser/SerializedPlanParser.cpp b/cpp-ch/local-engine/Parser/SerializedPlanParser.cpp
@@ -1105,6 +1105,7 @@ const ActionsDAG::Node * SerializedPlanParser::parseExpression(ActionsDAG& actio
                 throw Exception(ErrorCodes::BAD_ARGUMENTS, "Doesn't have type or input in cast node.");
             ActionsDAG::NodeRawConstPtrs args;
 
+            String cast_function = "CAST";
             const auto & input = rel.cast().input();
             args.emplace_back(parseExpression(actions_dag, input));
 
@@ -1144,10 +1145,15 @@ const ActionsDAG::Node * SerializedPlanParser::parseExpression(ActionsDAG& actio
                     /// Refer to https://github.com/apache/incubator-gluten/issues/4956
                     args[0] = toFunctionNode(actions_dag, "trim", {args[0]});
                 }
+                else if (isString(non_nullable_input_type) && substrait_type.has_bool_())
+                {
+                    /// cast(string to boolean)
+                    cast_function = "accurateCastOrNull";
+                }
 
                 /// Common process
                 args.emplace_back(addColumn(actions_dag, std::make_shared<DataTypeString>(), output_type->getName()));
-                function_node = toFunctionNode(actions_dag, "CAST", args);
+                function_node = toFunctionNode(actions_dag, cast_function, args);
             }
 
             actions_dag.addOrReplaceInOutputs(*function_node);