support creating hive table with DataFrameWriter and Catalog

Author: cloud-fan

sql/core/src/main/scala/org/apache/spark/sql/DataFrameWriter.scala

@@ -26,7 +26,6 @@ import org.apache.spark.sql.catalyst.TableIdentifier
 import org.apache.spark.sql.catalyst.analysis.{EliminateSubqueryAliases, UnresolvedRelation}
 import org.apache.spark.sql.catalyst.catalog.{BucketSpec, CatalogTable, CatalogTableType}
 import org.apache.spark.sql.catalyst.plans.logical.InsertIntoTable
-import org.apache.spark.sql.execution.command.DDLUtils
 import org.apache.spark.sql.execution.datasources.{CreateTable, DataSource, LogicalRelation}
 import org.apache.spark.sql.sources.BaseRelation
 import org.apache.spark.sql.types.StructType
@@ -361,10 +360,6 @@ final class DataFrameWriter[T] private[sql](ds: Dataset[T]) {
   }
 
   private def saveAsTable(tableIdent: TableIdentifier): Unit = {
-    if (source.toLowerCase == DDLUtils.HIVE_PROVIDER) {
-      throw new AnalysisException("Cannot create hive serde table with saveAsTable API")
-    }
-
     val catalog = df.sparkSession.sessionState.catalog
     val tableExists = catalog.tableExists(tableIdent)
     val db = tableIdent.database.getOrElse(catalog.getCurrentDatabase)
@@ -385,6 +380,8 @@ final class DataFrameWriter[T] private[sql](ds: Dataset[T]) {
         }
         EliminateSubqueryAliases(catalog.lookupRelation(tableIdentWithDB)) match {
           // Only do the check if the table is a data source table (the relation is a BaseRelation).
+          // TODO(cloud-fan): also check hive table relation here when we support overwrite mode
+          // for creating hive tables.
           case LogicalRelation(dest: BaseRelation, _, _) if srcRelations.contains(dest) =>
             throw new AnalysisException(
               s"Cannot overwrite table $tableName that is also being read from")

sql/core/src/main/scala/org/apache/spark/sql/internal/CatalogImpl.scala

@@ -347,10 +347,6 @@ class CatalogImpl(sparkSession: SparkSession) extends Catalog {
       source: String,
       schema: StructType,
       options: Map[String, String]): DataFrame = {
-    if (source.toLowerCase == "hive") {
-      throw new AnalysisException("Cannot create hive serde table with createExternalTable API.")
-    }
-
     val tableIdent = sparkSession.sessionState.sqlParser.parseTableIdentifier(tableName)
     val tableDesc = CatalogTable(
       identifier = tableIdent,

sql/hive/src/test/scala/org/apache/spark/sql/hive/execution/HiveDDLSuite.scala

@@ -33,6 +33,7 @@ import org.apache.spark.sql.hive.test.TestHiveSingleton
 import org.apache.spark.sql.internal.SQLConf
 import org.apache.spark.sql.internal.StaticSQLConf.CATALOG_IMPLEMENTATION
 import org.apache.spark.sql.test.SQLTestUtils
+import org.apache.spark.sql.types.StructType
 
 class HiveDDLSuite
   extends QueryTest with SQLTestUtils with TestHiveSingleton with BeforeAndAfterEach {
@@ -1289,4 +1290,66 @@ class HiveDDLSuite
       }
     }
   }
+
+  test("create hive serde table with Catalog") {
+    withTable("t") {
+      withTempDir { dir =>
+        val df = spark.catalog.createExternalTable(
+          "t",
+          "hive",
+          new StructType().add("i", "int"),
+          Map("path" -> dir.getCanonicalPath, "fileFormat" -> "parquet"))
+        assert(df.collect().isEmpty)
+
+        val table = spark.sessionState.catalog.getTableMetadata(TableIdentifier("t"))
+        assert(DDLUtils.isHiveTable(table))
+        assert(table.storage.inputFormat ==
+          Some("org.apache.hadoop.hive.ql.io.parquet.MapredParquetInputFormat"))
+        assert(table.storage.outputFormat ==
+          Some("org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat"))
+        assert(table.storage.serde ==
+          Some("org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe"))
+
+        sql("INSERT INTO t SELECT 1")
+        checkAnswer(spark.table("t"), Row(1))
+      }
+    }
+  }
+
+  test("create hive serde table with DataFrameWriter.saveAsTable") {
+    withTable("t", "t2") {
+      Seq(1 -> "a").toDF("i", "j")
+        .write.format("hive").option("fileFormat", "avro").saveAsTable("t")
+      checkAnswer(spark.table("t"), Row(1, "a"))
+
+      val table = spark.sessionState.catalog.getTableMetadata(TableIdentifier("t"))
+      assert(DDLUtils.isHiveTable(table))
+      assert(table.storage.inputFormat ==
+        Some("org.apache.hadoop.hive.ql.io.avro.AvroContainerInputFormat"))
+      assert(table.storage.outputFormat ==
+        Some("org.apache.hadoop.hive.ql.io.avro.AvroContainerOutputFormat"))
+      assert(table.storage.serde ==
+        Some("org.apache.hadoop.hive.serde2.avro.AvroSerDe"))
+
+      sql("INSERT INTO t SELECT 2, 'b'")
+      checkAnswer(spark.table("t"), Row(1, "a") :: Row(2, "b") :: Nil)
+
+      val e = intercept[AnalysisException] {
+        Seq(1 -> "a").toDF("i", "j").write.format("hive").partitionBy("i").saveAsTable("t2")
+      }
+      assert(e.message.contains("A Create Table As Select (CTAS) statement is not allowed " +
+        "to create a partitioned table using Hive"))
+
+      val e2 = intercept[AnalysisException] {
+        Seq(1 -> "a").toDF("i", "j").write.format("hive").bucketBy(4, "i").saveAsTable("t2")
+      }
+      assert(e2.message.contains("Creating bucketed Hive serde table is not supported yet"))
+
+      val e3 = intercept[AnalysisException] {
+        spark.table("t").write.format("hive").mode("overwrite").saveAsTable("t")
+      }
+      assert(e3.message.contains(
+        "CTAS for hive serde tables does not support append or overwrite semantics"))
+    }
+  }
 }