ParquetFileFormat¶

[[shortName]] ParquetFileFormat is the FileFormat for parquet data source (i.e. registers itself to handle files in parquet format and converts them to Spark SQL rows).

NOTE: parquet is the default data source format in Spark SQL.

// All the following queries are equivalent
// schema has to be specified manually
import org.apache.spark.sql.types.StructType
val schema = StructType($"id".int :: Nil)

spark.read.schema(schema).format("parquet").load("parquet-datasets")

// The above is equivalent to the following shortcut
// Implicitly does format("parquet").load
spark.read.schema(schema).parquet("parquet-datasets")

// parquet is the default data source format
spark.read.schema(schema).load("parquet-datasets")

[[isSplitable]] ParquetFileFormat is splitable, i.e. FIXME

[[supportBatch]] ParquetFileFormat supports vectorized parquet decoding in whole-stage code generation when all of the following hold:

• spark.sql.parquet.enableVectorizedReader configuration property is enabled

• spark.sql.codegen.wholeStage internal configuration property is enabled

• The number of fields in the schema is at most spark.sql.codegen.maxFields internal configuration property

• All the fields in the output schema are of AtomicType

ParquetFileFormat supports filter predicate push-down optimization (via <>) as per the following <>.

[[ParquetFilters]] .Spark Data Source Filters to Parquet Filter Predicates Conversions (aka ParquetFilters.createFilter) [cols="1m,2",options="header",width="100%"] |=== | Data Source Filter | Parquet FilterPredicate

| IsNull | [[IsNull]] FilterApi.eq

| IsNotNull | [[IsNotNull]] FilterApi.notEq

| EqualTo | [[EqualTo]] FilterApi.eq

| Not EqualTo | [[NotEqualTo]] FilterApi.notEq

| EqualNullSafe | [[EqualNullSafe]] FilterApi.eq

| Not EqualNullSafe | [[NotEqualNullSafe]] FilterApi.notEq

| LessThan | [[LessThan]] FilterApi.lt

| LessThanOrEqual | [[LessThanOrEqual]] FilterApi.ltEq

| GreaterThan | [[GreaterThan]] FilterApi.gt

| GreaterThanOrEqual | [[GreaterThanOrEqual]] FilterApi.gtEq

| And | [[And]] FilterApi.and

| Or | [[Or]] FilterApi.or

| No | [[Not]] FilterApi.not |===

[[logging]] [TIP] ==== Enable ALL logging level for org.apache.spark.sql.execution.datasources.parquet.ParquetFileFormat logger to see what happens inside.

Add the following line to conf/log4j.properties:

log4j.logger.org.apache.spark.sql.execution.datasources.parquet.ParquetFileFormat=ALL

Refer to <>.¶

=== [[prepareWrite]] Preparing Write Job -- prepareWrite Method

[source, scala]¶

prepareWrite( sparkSession: SparkSession, job: Job, options: Map[String, String], dataSchema: StructType): OutputWriterFactory

prepareWrite...FIXME

prepareWrite is part of the FileFormat abstraction.

=== [[inferSchema]] inferSchema Method

[source, scala]¶

inferSchema( sparkSession: SparkSession, parameters: Map[String, String], files: Seq[FileStatus]): Option[StructType]

inferSchema...FIXME

inferSchema is part of the FileFormat abstraction.

=== [[vectorTypes]] vectorTypes Method

[source, scala]¶

vectorTypes( requiredSchema: StructType, partitionSchema: StructType, sqlConf: SQLConf): Option[Seq[String]]

vectorTypes creates a collection of the names of OffHeapColumnVector or OnHeapColumnVector when spark.sql.columnVector.offheap.enabled property is enabled or disabled, respectively.

The size of the collection are all the fields of the given requiredSchema and partitionSchema schemas.

vectorTypes is part of the FileFormat abstraction.

=== [[buildReaderWithPartitionValues]] Building Data Reader With Partition Column Values Appended -- buildReaderWithPartitionValues Method

[source, scala]¶

buildReaderWithPartitionValues( sparkSession: SparkSession, dataSchema: StructType, partitionSchema: StructType, requiredSchema: StructType, filters: Seq[Filter], options: Map[String, String], hadoopConf: Configuration): (PartitionedFile) => Iterator[InternalRow]

buildReaderWithPartitionValues is part of the FileFormat abstraction.

buildReaderWithPartitionValues sets the <> in the input hadoopConf.

[[options]] .Hadoop Configuration Options [cols="1m,3",options="header",width="100%"] |=== | Name | Value

| parquet.read.support.class | [[parquet.read.support.class]] ParquetReadSupport

| org.apache.spark.sql.parquet.row.requested_schema | [[org.apache.spark.sql.parquet.row.requested_schema]] JSON representation of requiredSchema

| org.apache.spark.sql.parquet.row.attributes | [[org.apache.spark.sql.parquet.row.attributes]] JSON representation of requiredSchema

| spark.sql.session.timeZone | [[spark.sql.session.timeZone]] spark.sql.session.timeZone

| spark.sql.parquet.binaryAsString | [[spark.sql.parquet.binaryAsString]] spark.sql.parquet.binaryAsString

| spark.sql.parquet.int96AsTimestamp | [[spark.sql.parquet.int96AsTimestamp]] spark.sql.parquet.int96AsTimestamp

|===

buildReaderWithPartitionValues requests ParquetWriteSupport to setSchema.

buildReaderWithPartitionValues tries to push filters down to create a Parquet FilterPredicate (aka pushed).

With spark.sql.parquet.filterPushdown configuration property enabled, buildReaderWithPartitionValues takes the input Spark data source filters and converts them to Parquet filter predicates if possible (as described in the <>). Otherwise, the Parquet filter predicate is not specified.

buildReaderWithPartitionValues broadcasts the input hadoopConf Hadoop Configuration.

In the end, buildReaderWithPartitionValues gives a function that takes a PartitionedFile and does the following:

. Creates a Hadoop FileSplit for the input PartitionedFile

. Creates a Parquet ParquetInputSplit for the Hadoop FileSplit created

. Gets the broadcast Hadoop Configuration

. Creates a flag that says whether to apply timezone conversions to int96 timestamps or not (aka convertTz)

. Creates a Hadoop TaskAttemptContextImpl (with the broadcast Hadoop Configuration and a Hadoop TaskAttemptID for a map task)

. Sets the Parquet FilterPredicate (only when spark.sql.parquet.filterPushdown configuration property is enabled and it is by default)

The function then branches off on whether Parquet vectorized reader is enabled or not.

With Parquet vectorized reader enabled, the function does the following:

• Creates a VectorizedParquetRecordReader and a RecordReaderIterator

• Requests VectorizedParquetRecordReader to initialize (with the Parquet ParquetInputSplit and the Hadoop TaskAttemptContextImpl)

• Prints out the following DEBUG message to the logs:

  Appending [partitionSchema] [partitionValues]
  

• Requests VectorizedParquetRecordReader to initBatch

• (only with <> enabled) Requests VectorizedParquetRecordReader to enableReturningBatches

• In the end, the function gives the RecordReaderIterator (over the VectorizedParquetRecordReader) as the Iterator[InternalRow]

With Parquet vectorized reader disabled, the function does the following:

• FIXME (since Parquet vectorized reader is enabled by default it's of less interest currently)

=== [[mergeSchemasInParallel]] mergeSchemasInParallel Method

[source, scala]¶

mergeSchemasInParallel( filesToTouch: Seq[FileStatus], sparkSession: SparkSession): Option[StructType]

mergeSchemasInParallel...FIXME

NOTE: mergeSchemasInParallel is used when...FIXME