Enable map_concat function

Author: rui-mo

backends-clickhouse/src/main/scala/org/apache/gluten/utils/CHExpressionUtil.scala

@@ -204,6 +204,7 @@ object CHExpressionUtil {
     STACK -> DefaultValidator(),
     RAISE_ERROR -> DefaultValidator(),
     WIDTH_BUCKET -> DefaultValidator(),
-    MAKE_DATE -> DefaultValidator()
+    MAKE_DATE -> DefaultValidator(),
+    MAP_CONCAT -> DefaultValidator()
   )
 }

backends-velox/src/test/scala/org/apache/gluten/execution/ScalarFunctionsValidateSuite.scala

@@ -578,6 +578,24 @@ abstract class ScalarFunctionsValidateSuite extends FunctionsValidateSuite {
     }
   }
 
+  test("map_concat") {
+    withTempPath {
+      path =>
+        Seq(
+          Map[String, Int]("a" -> 1, "b" -> 2),
+          Map[String, Int]("a" -> 2, "b" -> 3),
+          null
+        )
+          .toDF("m")
+          .write
+          .parquet(path.getCanonicalPath)
+        spark.read.parquet(path.getCanonicalPath).createOrReplaceTempView("map_tbl")
+        runQueryAndCompare("select map_concat(m, map('c', 4)) from map_tbl") {
+          checkGlutenOperatorMatch[ProjectExecTransformer]
+        }
+    }
+  }
+
   test("map_filter") {
     withTempPath {
       path =>

cpp/core/config/GlutenConfig.h

@@ -75,6 +75,7 @@ const std::string kSparkRedactionString = "*********(redacted)";
 
 const std::string kSparkLegacyTimeParserPolicy = "spark.sql.legacy.timeParserPolicy";
 const std::string kShuffleFileBufferSize = "spark.shuffle.file.buffer";
+const std::string kSparkMapKeyDedupPolicy = "spark.sql.mapKeyDedupPolicy";
 
 std::unordered_map<std::string, std::string>
 parseConfMap(JNIEnv* env, const uint8_t* planData, const int32_t planDataLength);

cpp/velox/compute/WholeStageResultIterator.cc

@@ -566,6 +566,12 @@ std::unordered_map<std::string, std::string> WholeStageResultIterator::getQueryC
       configs[velox::core::QueryConfig::kSparkLegacyDateFormatter] = "false";
     }
 
+    if (veloxCfg_->get<std::string>(kSparkMapKeyDedupPolicy, "") == "EXCEPTION") {
+      configs[velox::core::QueryConfig::kSparkThrowExceptionOnDuplicateMapKeys] = "true";
+    } else {
+      configs[velox::core::QueryConfig::kSparkThrowExceptionOnDuplicateMapKeys] = "false";
+    }
+
     const auto setIfExists = [&](const std::string& glutenKey, const std::string& veloxKey) {
       const auto valueOptional = veloxCfg_->get<std::string>(glutenKey);
       if (valueOptional.hasValue()) {

gluten-substrait/src/main/scala/org/apache/gluten/expression/ExpressionMappings.scala

@@ -259,6 +259,7 @@ object ExpressionMappings {
     // Map functions
     Sig[CreateMap](CREATE_MAP),
     Sig[GetMapValue](GET_MAP_VALUE),
+    Sig[MapConcat](MAP_CONCAT),
     Sig[MapKeys](MAP_KEYS),
     Sig[MapValues](MAP_VALUES),
     Sig[MapFromArrays](MAP_FROM_ARRAYS),

shims/common/src/main/scala/org/apache/gluten/config/GlutenConfig.scala

@@ -506,7 +506,8 @@ object GlutenConfig {
       (
         GLUTEN_COLUMNAR_TO_ROW_MEM_THRESHOLD.key,
         GLUTEN_COLUMNAR_TO_ROW_MEM_THRESHOLD.defaultValue.get.toString),
-      (SPARK_SHUFFLE_SPILL_COMPRESS, SPARK_SHUFFLE_SPILL_COMPRESS_DEFAULT.toString)
+      (SPARK_SHUFFLE_SPILL_COMPRESS, SPARK_SHUFFLE_SPILL_COMPRESS_DEFAULT.toString),
+      (SQLConf.MAP_KEY_DEDUP_POLICY.key, SQLConf.MAP_KEY_DEDUP_POLICY.defaultValueString)
     )
     keyWithDefault.forEach(e => nativeConfMap.put(e._1, conf.getOrElse(e._1, e._2)))
     GlutenConfigUtil.mapByteConfValue(

shims/common/src/main/scala/org/apache/gluten/expression/ExpressionNames.scala

@@ -282,6 +282,7 @@ object ExpressionNames {
   // Map functions
   final val CREATE_MAP = "map"
   final val GET_MAP_VALUE = "get_map_value"
+  final val MAP_CONCAT = "map_concat"
   final val MAP_KEYS = "map_keys"
   final val MAP_VALUES = "map_values"
   final val MAP_FROM_ARRAYS = "map_from_arrays"