[SPARK-6538][SQL] Add missing nullable Metastore fields when merging a Parquet schema

Opening to replace #5188.

When Spark SQL infers a schema for a DataFrame, it will take the union of all field types present in the structured source data (e.g. an RDD of JSON data). When the source data for a row doesn't define a particular field on the DataFrame's schema, a null value will simply be assumed for this field. This workflow makes it very easy to construct tables and query over a set of structured data with a nonuniform schema. However, this behavior is not consistent in some cases when dealing with Parquet files and an external table managed by an external Hive metastore.

In our particular usecase, we use Spark Streaming to parse and transform our input data and then apply a window function to save an arbitrary-sized batch of data as a Parquet file, which itself will be added as a partition to an external Hive table via an *"ALTER TABLE... ADD PARTITION..."* statement. Since our input data is nonuniform, it is expected that not every partition batch will contain every field present in the table's schema obtained from the Hive metastore. As such, we expect that the schema of some of our Parquet files may not contain the same set fields present in the full metastore schema.

In such cases, it seems natural that Spark SQL would simply assume null values for any missing fields in the partition's Parquet file, assuming these fields are specified as nullable by the metastore schema. This is not the case in the current implementation of ParquetRelation2. The **mergeMetastoreParquetSchema()** method used to reconcile differences between a Parquet file's schema and a schema retrieved from the Hive metastore will raise an exception if the Parquet file doesn't match the same set of fields specified by the metastore.

This pull requests alters the behavior of **mergeMetastoreParquetSchema()** by having it first add any nullable fields from the metastore schema to the Parquet file schema if they aren't already present there.

Author: Adam Budde <[email protected]>

Closes #5214 from budde/nullable-fields and squashes the following commits:

a52d378 [Adam Budde] Refactor ParquetSchemaSuite.scala for cases now permitted by SPARK-6471 and SPARK-6538
9041bfa [Adam Budde] Add missing nullable Metastore fields when merging a Parquet schema

(cherry picked from commit 5909f09)
Signed-off-by: Cheng Lian <[email protected]>

Author: Adam Budde

sql/core/src/main/scala/org/apache/spark/sql/parquet/newParquet.scala

@@ -765,12 +765,14 @@ private[sql] object ParquetRelation2 extends Logging {
          |${parquetSchema.prettyJson}
        """.stripMargin
 
-    assert(metastoreSchema.size <= parquetSchema.size, schemaConflictMessage)
+    val mergedParquetSchema = mergeMissingNullableFields(metastoreSchema, parquetSchema)
+
+    assert(metastoreSchema.size <= mergedParquetSchema.size, schemaConflictMessage)
 
     val ordinalMap = metastoreSchema.zipWithIndex.map {
       case (field, index) => field.name.toLowerCase -> index
     }.toMap
-    val reorderedParquetSchema = parquetSchema.sortBy(f => 
+    val reorderedParquetSchema = mergedParquetSchema.sortBy(f => 
       ordinalMap.getOrElse(f.name.toLowerCase, metastoreSchema.size + 1))
 
     StructType(metastoreSchema.zip(reorderedParquetSchema).map {
@@ -782,6 +784,32 @@ private[sql] object ParquetRelation2 extends Logging {
     })
   }
 
+  /**
+   * Returns the original schema from the Parquet file with any missing nullable fields from the
+   * Hive Metastore schema merged in.
+   *
+   * When constructing a DataFrame from a collection of structured data, the resulting object has
+   * a schema corresponding to the union of the fields present in each element of the collection.
+   * Spark SQL simply assigns a null value to any field that isn't present for a particular row.
+   * In some cases, it is possible that a given table partition stored as a Parquet file doesn't
+   * contain a particular nullable field in its schema despite that field being present in the
+   * table schema obtained from the Hive Metastore. This method returns a schema representing the
+   * Parquet file schema along with any additional nullable fields from the Metastore schema
+   * merged in.
+   */
+  private[parquet] def mergeMissingNullableFields(
+      metastoreSchema: StructType,
+      parquetSchema: StructType): StructType = {
+    val fieldMap = metastoreSchema.map(f => f.name.toLowerCase -> f).toMap
+    val missingFields = metastoreSchema
+      .map(_.name.toLowerCase)
+      .diff(parquetSchema.map(_.name.toLowerCase))
+      .map(fieldMap(_))
+      .filter(_.nullable)
+    StructType(parquetSchema ++ missingFields)
+  }
+
+
   // TODO Data source implementations shouldn't touch Catalyst types (`Literal`).
   // However, we are already using Catalyst expressions for partition pruning and predicate
   // push-down here...

sql/core/src/test/scala/org/apache/spark/sql/parquet/ParquetSchemaSuite.scala

@@ -226,22 +226,54 @@ class ParquetSchemaSuite extends FunSuite with ParquetTest {
           StructField("UPPERCase", IntegerType, nullable = true))))
     }
 
-    // Conflicting field count
+    // Metastore schema contains additional non-nullable fields.
     assert(intercept[Throwable] {
       ParquetRelation2.mergeMetastoreParquetSchema(
         StructType(Seq(
           StructField("uppercase", DoubleType, nullable = false),
-          StructField("lowerCase", BinaryType))),
+          StructField("lowerCase", BinaryType, nullable = false))),
 
         StructType(Seq(
           StructField("UPPERCase", IntegerType, nullable = true))))
     }.getMessage.contains("detected conflicting schemas"))
 
-    // Conflicting field names
+    // Conflicting non-nullable field names
     intercept[Throwable] {
       ParquetRelation2.mergeMetastoreParquetSchema(
-        StructType(Seq(StructField("lower", StringType))),
+        StructType(Seq(StructField("lower", StringType, nullable = false))),
         StructType(Seq(StructField("lowerCase", BinaryType))))
     }
   }
+
+  test("merge missing nullable fields from Metastore schema") {
+    // Standard case: Metastore schema contains additional nullable fields not present
+    // in the Parquet file schema.
+    assertResult(
+      StructType(Seq(
+        StructField("firstField", StringType, nullable = true),
+        StructField("secondField", StringType, nullable = true),
+        StructField("thirdfield", StringType, nullable = true)))) {
+      ParquetRelation2.mergeMetastoreParquetSchema(
+        StructType(Seq(
+          StructField("firstfield", StringType, nullable = true),
+          StructField("secondfield", StringType, nullable = true),
+          StructField("thirdfield", StringType, nullable = true))),
+        StructType(Seq(
+          StructField("firstField", StringType, nullable = true),
+          StructField("secondField", StringType, nullable = true))))
+    }
+
+    // Merge should fail if the Metastore contains any additional fields that are not
+    // nullable.
+    assert(intercept[Throwable] {
+      ParquetRelation2.mergeMetastoreParquetSchema(
+        StructType(Seq(
+          StructField("firstfield", StringType, nullable = true),
+          StructField("secondfield", StringType, nullable = true),
+          StructField("thirdfield", StringType, nullable = false))),
+        StructType(Seq(
+          StructField("firstField", StringType, nullable = true),
+          StructField("secondField", StringType, nullable = true))))
+    }.getMessage.contains("detected conflicting schemas"))
+  }
 }