com.nvidia.spark.rapids.ParquetCachedBatchSerializer.scala Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of rapids-4-spark_2.12 Show documentation
Show all versions of rapids-4-spark_2.12 Show documentation
Creates the distribution package of the RAPIDS plugin for Apache Spark
/*
* Copyright (c) 2020-2023, NVIDIA CORPORATION.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package com.nvidia.spark.rapids
import java.io.{InputStream, IOException}
import java.nio.ByteBuffer
import scala.collection.JavaConverters._
import scala.collection.mutable
import scala.collection.mutable.ListBuffer
import ai.rapids.cudf._
import ai.rapids.cudf.ParquetWriterOptions.StatisticsFrequency
import com.nvidia.spark.GpuCachedBatchSerializer
import com.nvidia.spark.rapids.Arm.withResource
import com.nvidia.spark.rapids.GpuColumnVector.GpuColumnarBatchBuilder
import com.nvidia.spark.rapids.RapidsPluginImplicits._
import com.nvidia.spark.rapids.ScalableTaskCompletion.onTaskCompletion
import com.nvidia.spark.rapids.shims.{LegacyBehaviorPolicyShim, ParquetFieldIdShims, ParquetLegacyNanoAsLongShims, ParquetTimestampNTZShims, SparkShimImpl}
import org.apache.commons.io.output.ByteArrayOutputStream
import org.apache.hadoop.conf.Configuration
import org.apache.hadoop.mapreduce.RecordWriter
import org.apache.parquet.{HadoopReadOptions, ParquetReadOptions}
import org.apache.parquet.column.ParquetProperties
import org.apache.parquet.hadoop.{CodecFactory, MemoryManager, ParquetFileReader, ParquetFileWriter, ParquetInputFormat, ParquetOutputFormat, ParquetRecordWriter, ParquetWriter}
import org.apache.parquet.hadoop.ParquetFileWriter.Mode
import org.apache.parquet.hadoop.api.WriteSupport
import org.apache.parquet.hadoop.metadata.CompressionCodecName
import org.apache.parquet.io.{DelegatingPositionOutputStream, DelegatingSeekableInputStream, InputFile, OutputFile, PositionOutputStream}
import org.apache.spark.TaskContext
import org.apache.spark.broadcast.Broadcast
import org.apache.spark.rdd.RDD
import org.apache.spark.sql.SparkSession
import org.apache.spark.sql.catalyst.InternalRow
import org.apache.spark.sql.catalyst.expressions.{Attribute, AttributeReference, Expression, GenericInternalRow, SpecializedGetters, UnsafeRow}
import org.apache.spark.sql.catalyst.expressions.codegen.GenerateUnsafeProjection
import org.apache.spark.sql.catalyst.util.{ArrayBasedMapData, ArrayData, GenericArrayData, MapData}
import org.apache.spark.sql.columnar.CachedBatch
import org.apache.spark.sql.execution.datasources.parquet.{ParquetReadSupport, ParquetToSparkSchemaConverter, ParquetWriteSupport, ShimCurrentBatchIterator}
import org.apache.spark.sql.execution.datasources.parquet.rapids.ParquetRecordMaterializer
import org.apache.spark.sql.internal.SQLConf
import org.apache.spark.sql.rapids.PCBSSchemaHelper
import org.apache.spark.sql.types._
import org.apache.spark.sql.vectorized.ColumnarBatch
import org.apache.spark.storage.StorageLevel
import org.apache.spark.unsafe.types.CalendarInterval
/**
* copied from Spark org.apache.spark.util.ByteBufferInputStream
*/
private class ByteBufferInputStream(private var buffer: ByteBuffer)
extends InputStream {
override def read(): Int = {
if (buffer == null || buffer.remaining() == 0) {
cleanUp()
-1
} else {
buffer.get() & 0xFF
}
}
override def read(dest: Array[Byte]): Int = {
read(dest, 0, dest.length)
}
override def read(dest: Array[Byte], offset: Int, length: Int): Int = {
if (buffer == null || buffer.remaining() == 0) {
cleanUp()
-1
} else {
val amountToGet = math.min(buffer.remaining(), length)
buffer.get(dest, offset, amountToGet)
amountToGet
}
}
override def skip(bytes: Long): Long = {
if (buffer != null) {
val amountToSkip = math.min(bytes, buffer.remaining).toInt
buffer.position(buffer.position() + amountToSkip)
if (buffer.remaining() == 0) {
cleanUp()
}
amountToSkip
} else {
0L
}
}
/**
* Clean up the buffer, and potentially dispose of it using StorageUtils.dispose().
*/
private def cleanUp(): Unit = {
if (buffer != null) {
buffer = null
}
}
}
class ByteArrayInputFile(buff: Array[Byte]) extends InputFile {
override def getLength: Long = buff.length
override def newStream(): org.apache.parquet.io.SeekableInputStream = {
val byteBuffer = ByteBuffer.wrap(buff)
new DelegatingSeekableInputStream(new ByteBufferInputStream(byteBuffer)) {
override def getPos: Long = byteBuffer.position()
override def seek(newPos: Long): Unit = {
if (newPos > Int.MaxValue || newPos < Int.MinValue) {
throw new IllegalStateException("seek value is out of supported range " + newPos)
}
byteBuffer.position(newPos.toInt)
}
}
}
}
private object ByteArrayOutputFile {
val BLOCK_SIZE: Int = 32 * 1024 * 1024 // 32M
}
private class ByteArrayOutputFile(stream: ByteArrayOutputStream) extends OutputFile {
override def create(blockSizeHint: Long): PositionOutputStream = {
new DelegatingPositionOutputStream(stream) {
var pos = 0
override def getPos: Long = pos
override def write(b: Int): Unit = {
super.write(b)
pos += Integer.BYTES
}
override def write(b: Array[Byte]): Unit = {
super.write(b)
pos += b.length
}
override def write(b: Array[Byte], off: Int, len: Int): Unit = {
super.write(b, off, len)
pos += len
}
}
}
override def createOrOverwrite(blockSizeHint: Long): PositionOutputStream =
throw new UnsupportedOperationException("Don't need to overwrite")
override def supportsBlockSize(): Boolean = true
override def defaultBlockSize(): Long = ByteArrayOutputFile.BLOCK_SIZE
}
private class ParquetBufferConsumer(val numRows: Int) extends HostBufferConsumer with
AutoCloseable {
@transient private[this] val offHeapBuffers = mutable.Queue[(HostMemoryBuffer, Long)]()
private var buffer: Array[Byte] = _
override def handleBuffer(buffer: HostMemoryBuffer, len: Long): Unit = {
offHeapBuffers += Tuple2(buffer, len)
}
def getBuffer: Array[Byte] = {
if (buffer == null) {
writeBuffers()
}
buffer
}
def close(): Unit = {
if (buffer == null) {
writeBuffers()
}
}
private def writeBuffers(): Unit = {
val toProcess = offHeapBuffers.dequeueAll(_ => true)
// We are making sure the input is smaller than 2gb so the parquet written should never be more
// than Int.MAX_SIZE.
val bytes = toProcess.map(_._2).sum
// for now assert bytes are less than Int.MaxValue
assert(bytes <= Int.MaxValue)
buffer = new Array(bytes.toInt)
try {
var offset: Int = 0
toProcess.foreach(ops => {
val origBuffer = ops._1
val len = ops._2.toInt
origBuffer.asByteBuffer().get(buffer, offset, len)
offset = offset + len
})
} finally {
toProcess.map(_._1).safeClose()
}
}
}
private object ParquetCachedBatch {
def apply(parquetBuff: ParquetBufferConsumer): ParquetCachedBatch = {
new ParquetCachedBatch(parquetBuff.numRows, parquetBuff.getBuffer)
}
}
case class ParquetCachedBatch(
numRows: Int,
buffer: Array[Byte]) extends CachedBatch {
override def sizeInBytes: Long = buffer.length
}
/**
* Spark wants the producer to close the batch. We have a listener in this iterator that will close
* the batch after the task is completed
*/
private case class CloseableColumnBatchIterator(iter: Iterator[ColumnarBatch]) extends
Iterator[ColumnarBatch] {
var cb: ColumnarBatch = _
private def closeCurrentBatch(): Unit = {
if (cb != null) {
cb.close()
cb = null
}
}
// Don't install the callback if in a unit test
Option(TaskContext.get()).foreach { tc =>
onTaskCompletion(tc) {
closeCurrentBatch()
}
}
override def hasNext: Boolean = iter.hasNext
override def next(): ColumnarBatch = {
closeCurrentBatch()
cb = iter.next()
cb
}
}
/**
* This class assumes, the data is Columnar and the plugin is on.
* Note, this class should not be referenced directly in source code.
* It should be loaded by reflection using ShimLoader.newInstanceOf, see ./docs/dev/shims.md
*/
protected class ParquetCachedBatchSerializer extends GpuCachedBatchSerializer {
override def supportsColumnarInput(schema: Seq[Attribute]): Boolean = true
override def supportsColumnarOutput(schema: StructType): Boolean = schema.fields.forall { f =>
// only check spark b/c if we are on the GPU then we will be calling the gpu method regardless
PCBSSchemaHelper.isTypeSupportedByColumnarSparkParquetWriter(f.dataType) ||
f.dataType == DataTypes.NullType
}
def isSchemaSupportedByCudf(schema: Seq[Attribute]): Boolean = {
schema.forall(field => isSupportedByCudf(field.dataType))
}
def isSupportedByCudf(dataType: DataType): Boolean = {
dataType match {
case a: ArrayType => isSupportedByCudf(a.elementType)
case s: StructType => s.forall(field => isSupportedByCudf(field.dataType))
case m: MapType => isSupportedByCudf(m.keyType) && isSupportedByCudf(m.valueType)
case _ => GpuColumnVector.isNonNestedSupportedType(dataType)
}
}
/**
* Convert an `RDD[ColumnarBatch]` into an `RDD[CachedBatch]` in preparation for caching the data.
* This method uses Parquet Writer on the GPU to write the cached batch
*
* @param input the input `RDD` to be converted.
* @param schema the schema of the data being stored.
* @param storageLevel where the data will be stored.
* @param conf the config for the query.
* @return The data converted into a format more suitable for caching.
*/
override def convertColumnarBatchToCachedBatch(input: RDD[ColumnarBatch],
schema: Seq[Attribute],
storageLevel: StorageLevel,
conf: SQLConf): RDD[CachedBatch] = {
val rapidsConf = new RapidsConf(conf)
val useCompression = conf.useCompression
val bytesAllowedPerBatch = getBytesAllowedPerBatch(conf)
val (schemaWithUnambiguousNames, _) = getSupportedSchemaFromUnsupported(schema)
val structSchema = schemaWithUnambiguousNames.toStructType
if (rapidsConf.isSqlEnabled && rapidsConf.isSqlExecuteOnGPU &&
isSchemaSupportedByCudf(schema)) {
def putOnGpuIfNeeded(batch: ColumnarBatch): ColumnarBatch = {
if (!batch.column(0).isInstanceOf[GpuColumnVector]) {
// The input batch from CPU must NOT be closed, because the columns inside it
// will be reused, and Spark expects the producer to close its batches.
val numRows = batch.numRows()
withResource(new GpuColumnarBatchBuilder(structSchema, numRows)) { gcbBuilder =>
for (i <- 0 until batch.numCols()) {
gcbBuilder.copyColumnar(batch.column(i), i, numRows)
}
gcbBuilder.build(numRows)
}
} else {
batch
}
}
input.flatMap(batch => {
// If the structSchema is empty or the batch has no cols in that case return a batch with
// just rows.
if (batch.numCols() == 0 || structSchema.isEmpty) {
val rows = batch.numRows()
if (batch.numCols() > 0 && batch.column(0).isInstanceOf[GpuColumnVector]) {
batch.close()
}
List(ParquetCachedBatch(rows, new Array[Byte](0)))
} else {
withResource(putOnGpuIfNeeded(batch)) { gpuCB =>
compressColumnarBatchWithParquet(gpuCB, structSchema, schema.toStructType,
bytesAllowedPerBatch, useCompression)
}
}
})
} else {
val broadcastedConf = SparkSession.active.sparkContext.broadcast(conf.getAllConfs)
input.mapPartitions {
cbIter =>
new CachedBatchIteratorProducer[ColumnarBatch](cbIter, schemaWithUnambiguousNames, schema,
broadcastedConf).getColumnarBatchToCachedBatchIterator
}
}
}
private[rapids] def compressColumnarBatchWithParquet(
oldGpuCB: ColumnarBatch,
schema: StructType,
origSchema: StructType,
bytesAllowedPerBatch: Long,
useCompression: Boolean): List[ParquetCachedBatch] = {
if (oldGpuCB.numRows() == 0) {
// return empty List of ParquetCachedBatch
return List.empty[ParquetCachedBatch]
}
val estimatedRowSize = scala.Range(0, oldGpuCB.numCols()).map { idx =>
oldGpuCB.column(idx).asInstanceOf[GpuColumnVector]
.getBase.getDeviceMemorySize / oldGpuCB.numRows()
}.sum
val columns = for (i <- 0 until oldGpuCB.numCols()) yield {
val gpuVector = oldGpuCB.column(i).asInstanceOf[GpuColumnVector]
var dataType = origSchema(i).dataType
val v = ColumnCastUtil.ifTrueThenDeepConvertTypeAtoTypeB(gpuVector.getBase,
origSchema(i).dataType,
// we are checking for scale > 0 because cudf and spark refer to scales as opposites
// e.g. scale = -3 in Spark is scale = 3 in cudf
(_, cv) => cv.getType.isDecimalType && cv.getType.getScale > 0,
(_, cv) => {
if (cv.getType.isBackedByLong) {
dataType = LongType
cv.bitCastTo(DType.INT64)
} else {
dataType = IntegerType
cv.bitCastTo(DType.INT32)
}
}
)
GpuColumnVector.from(v, schema(i).dataType)
}
withResource(new ColumnarBatch(columns.toArray, oldGpuCB.numRows())) { gpuCB =>
val rowsAllowedInBatch = (bytesAllowedPerBatch / estimatedRowSize).toInt
val splitIndices = scala.Range(rowsAllowedInBatch, gpuCB.numRows(), rowsAllowedInBatch)
val buffers = new ListBuffer[ParquetCachedBatch]
if (splitIndices.nonEmpty) {
val splitVectors = new ListBuffer[Array[ColumnVector]]
try {
for (index <- 0 until gpuCB.numCols()) {
splitVectors +=
gpuCB.column(index).asInstanceOf[GpuColumnVector].getBase.split(splitIndices: _*)
}
// Splitting the table
// e.g. T0 = {col1, col2,...,coln} => split columns into 'm' cols =>
// T00= {splitCol1(0), splitCol2(0),...,splitColn(0)}
// T01= {splitCol1(1), splitCol2(1),...,splitColn(1)}
// ...
// T0m= {splitCol1(m), splitCol2(m),...,splitColn(m)}
def makeTableForIndex(i: Int): Table = {
val columns = splitVectors.indices.map(j => splitVectors(j)(i))
new Table(columns: _*)
}
for (i <- splitVectors.head.indices) {
withResource(makeTableForIndex(i)) { table =>
val buffer = writeTableToCachedBatch(table, schema, useCompression)
buffers += ParquetCachedBatch(buffer)
}
}
} finally {
splitVectors.foreach(array => array.safeClose())
}
} else {
withResource(GpuColumnVector.from(gpuCB)) { table =>
val buffer = writeTableToCachedBatch(table, schema, useCompression)
buffers += ParquetCachedBatch(buffer)
}
}
buffers.toList
}
}
def getParquetWriterOptions(
useCompression: Boolean,
schema: StructType): ParquetWriterOptions = {
val compressionType = if (useCompression) CompressionType.SNAPPY else CompressionType.NONE
SchemaUtils
.writerOptionsFromSchema(ParquetWriterOptions.builder(), schema, writeInt96 = false)
.withCompressionType(compressionType)
.withStatisticsFrequency(StatisticsFrequency.ROWGROUP).build()
}
private def writeTableToCachedBatch(
table: Table,
schema: StructType,
useCompression: Boolean): ParquetBufferConsumer = {
val buffer = new ParquetBufferConsumer(table.getRowCount.toInt)
val opts = getParquetWriterOptions(useCompression, schema)
withResource(Table.writeParquetChunked(opts, buffer)) { writer =>
writer.write(table)
}
buffer
}
/**
* This method decodes the CachedBatch leaving it on the GPU to avoid the extra copying back to
* the host
*
* @param input the cached batches that should be converted.
* @param cacheAttributes the attributes of the data in the batch.
* @param selectedAttributes the fields that should be loaded from the data and the order they
* should appear in the output batch.
* @param conf the configuration for the job.
* @return an RDD of the input cached batches transformed into the ColumnarBatch format.
*/
def gpuConvertCachedBatchToColumnarBatch(
input: RDD[CachedBatch],
cacheAttributes: Seq[Attribute],
selectedAttributes: Seq[Attribute],
conf: SQLConf): RDD[ColumnarBatch] = {
// optimize
val newSelectedAttributes = if (selectedAttributes.isEmpty) {
cacheAttributes
} else {
selectedAttributes
}
val (cachedSchemaWithNames, selectedSchemaWithNames) =
getSupportedSchemaFromUnsupported(cacheAttributes, newSelectedAttributes)
convertCachedBatchToColumnarInternal(
input,
cachedSchemaWithNames,
selectedSchemaWithNames,
newSelectedAttributes)
}
private def convertCachedBatchToColumnarInternal(
input: RDD[CachedBatch],
cacheAttributes: Seq[Attribute],
selectedAttributes: Seq[Attribute],
originalSelectedAttributes: Seq[Attribute]): RDD[ColumnarBatch] = {
val cbRdd: RDD[ColumnarBatch] = input.map {
case parquetCB: ParquetCachedBatch if parquetCB.sizeInBytes == 0 =>
// If the buffer is empty, we have cached a batch with no columns, we don't need to decode
// it, instead just return a ColumnarBatch with only rows
withResource(new GpuColumnarBatchBuilder(originalSelectedAttributes.toStructType,
parquetCB.numRows)) {
builder => builder.build(parquetCB.numRows)
}
case parquetCB: ParquetCachedBatch =>
val parquetOptions = ParquetOptions.builder()
.includeColumn(selectedAttributes.map(_.name).asJavaCollection).build()
val table = try {
Table.readParquet(parquetOptions, parquetCB.buffer, 0, parquetCB.sizeInBytes)
} catch {
case e: Exception =>
throw new IOException("Error when processing file " +
s"[range: 0-${parquetCB.sizeInBytes}]", e)
}
withResource(table) { table =>
withResource {
for (i <- 0 until table.getNumberOfColumns) yield {
ColumnCastUtil.ifTrueThenDeepConvertTypeAtoTypeB(table.getColumn(i),
originalSelectedAttributes(i).dataType,
(dataType, _) => dataType match {
case d: DecimalType if d.scale < 0 => true
case _ => false
},
(dataType, cv) => {
dataType match {
case d: DecimalType =>
withResource(cv.bitCastTo(DecimalUtil.createCudfDecimal(d))) {
_.copyToColumnVector()
}
case _ =>
throw new IllegalStateException("We don't cast any type besides Decimal " +
"with scale < 0")
}
}
)
}
} { col =>
withResource(new Table(col: _*)) { t =>
GpuColumnVector.from(t, originalSelectedAttributes.map(_.dataType).toArray)
}
}
}
case _ =>
throw new IllegalStateException("I don't know how to convert this batch")
}
cbRdd
}
private def getSelectedSchemaFromCachedSchema(
selectedAttributes: Seq[Attribute],
cacheAttributes: Seq[Attribute]): Seq[Attribute] = {
selectedAttributes.map {
a => cacheAttributes(cacheAttributes.map(_.exprId).indexOf(a.exprId))
}
}
/**
* Convert the cached data into a ColumnarBatch taking the result data back to the host
*
* @param input the cached batches that should be converted.
* @param cacheAttributes the attributes of the data in the batch.
* @param selectedAttributes the fields that should be loaded from the data and the order they
* should appear in the output batch.
* @param conf the configuration for the job.
* @return an RDD of the input cached batches transformed into the ColumnarBatch format.
*/
override def convertCachedBatchToColumnarBatch(input: RDD[CachedBatch],
cacheAttributes: Seq[Attribute],
selectedAttributes: Seq[Attribute],
conf: SQLConf): RDD[ColumnarBatch] = {
// optimize
val newSelectedAttributes = if (selectedAttributes.isEmpty) {
cacheAttributes
} else {
selectedAttributes
}
val rapidsConf = new RapidsConf(conf)
val (cachedSchemaWithNames, selectedSchemaWithNames) =
getSupportedSchemaFromUnsupported(cacheAttributes, newSelectedAttributes)
if (rapidsConf.isSqlEnabled && rapidsConf.isSqlExecuteOnGPU &&
isSchemaSupportedByCudf(cachedSchemaWithNames)) {
val batches = convertCachedBatchToColumnarInternal(input, cachedSchemaWithNames,
selectedSchemaWithNames, newSelectedAttributes)
val cbRdd = batches.map(batch => {
withResource(batch) { gpuBatch =>
val cols = GpuColumnVector.extractColumns(gpuBatch)
new ColumnarBatch(cols.safeMap(_.copyToHost()).toArray, gpuBatch.numRows())
}
})
cbRdd.mapPartitions(iter => CloseableColumnBatchIterator(iter))
} else {
val origSelectedAttributesWithUnambiguousNames =
sanitizeColumnNames(newSelectedAttributes, selectedSchemaWithNames)
val broadcastedConf = SparkSession.active.sparkContext.broadcast(conf.getAllConfs)
input.mapPartitions {
cbIter => {
new CachedBatchIteratorConsumer(cbIter, cachedSchemaWithNames, selectedSchemaWithNames,
cacheAttributes, origSelectedAttributesWithUnambiguousNames, broadcastedConf)
.getColumnBatchIterator
}
}
}
}
/**
* Convert the cached batch into `InternalRow`s.
*
* @param input the cached batches that should be converted.
* @param cacheAttributes the attributes of the data in the batch.
* @param selectedAttributes the field that should be loaded from the data and the order they
* should appear in the output rows.
* @param conf the configuration for the job.
* @return RDD of the rows that were stored in the cached batches.
*/
override def convertCachedBatchToInternalRow(
input: RDD[CachedBatch],
cacheAttributes: Seq[Attribute],
selectedAttributes: Seq[Attribute],
conf: SQLConf): RDD[InternalRow] = {
val (cachedSchemaWithNames, selectedSchemaWithNames) =
getSupportedSchemaFromUnsupported(cacheAttributes, selectedAttributes)
val newSelectedAttributes = sanitizeColumnNames(selectedAttributes, selectedSchemaWithNames)
val broadcastedConf = SparkSession.active.sparkContext.broadcast(conf.getAllConfs)
input.mapPartitions {
cbIter => {
new CachedBatchIteratorConsumer(cbIter, cachedSchemaWithNames, selectedSchemaWithNames,
cacheAttributes, newSelectedAttributes, broadcastedConf).getInternalRowIterator
}
}
}
/**
* Read the data from an input file and convert the result to rows. The data will not be
* transformed in any way. It will just be read.
*/
private class PcbFileToRowsIterator(inputFile: ByteArrayInputFile,
cacheAttributes: Seq[Attribute],
hadoopConf: Configuration,
options: ParquetReadOptions) extends Iterator[InternalRow] with AutoCloseable {
import org.apache.parquet.io.ColumnIOFactory
private[this] val parquetFileReader = ParquetFileReader.open(inputFile, options)
private[this] val parquetSchema = parquetFileReader.getFooter.getFileMetaData.getSchema
private[this] val columnIO = new ColumnIOFactory().getColumnIO(parquetSchema)
private[this] var pages: org.apache.parquet.column.page.PageReadStore = _
private[this] var rowsLeftInRowGroup: Long = _
private[this] var recordReader: org.apache.parquet.io.RecordReader[InternalRow] = _
// Prime the data with the first row group
goToNextRowGroup
private[this] def goToNextRowGroup: Unit = {
pages = parquetFileReader.readNextRowGroup()
if (pages == null) {
rowsLeftInRowGroup = 0
recordReader = null
} else {
rowsLeftInRowGroup = pages.getRowCount
recordReader = columnIO.getRecordReader(pages, new ParquetRecordMaterializer(parquetSchema,
cacheAttributes.toStructType,
new ParquetToSparkSchemaConverter(hadoopConf), None /*convertTz*/ ,
LegacyBehaviorPolicyShim.CORRECTED_STR))
}
}
override def hasNext: Boolean = pages != null
override def next(): InternalRow = {
if (!hasNext) {
throw new NoSuchElementException()
}
val ret = recordReader.read()
rowsLeftInRowGroup -= 1
if (rowsLeftInRowGroup <= 0) {
goToNextRowGroup
if (pages == null) {
close()
}
}
ret
}
override def close(): Unit = {
parquetFileReader.close()
}
}
object PcbToRowsIterator {
def apply(cbIter: Iterator[CachedBatch],
cacheAttributes: Seq[Attribute],
hadoopConf: Configuration,
options: ParquetReadOptions): Iterator[InternalRow] = {
new PcbToRowsIterator(cbIter, cacheAttributes, hadoopConf, options).flatten
}
}
/**
* Read the data from the CachedBatch iterator and return a row iterator for each batch.
*/
private class PcbToRowsIterator private(
cbIter: Iterator[CachedBatch],
cacheAttributes: Seq[Attribute],
hadoopConf: Configuration,
options: ParquetReadOptions) extends Iterator[Iterator[InternalRow]] {
val pcbIter = cbIter.map(_.asInstanceOf[ParquetCachedBatch])
override def hasNext: Boolean = pcbIter.hasNext
override def next(): Iterator[InternalRow] = {
val parquetCachedBatch = pcbIter.next()
if (parquetCachedBatch.sizeInBytes == 0) {
// Edge case where we have stored an empty parquet file possibly a dataframe with no
// columns
new ColumnarBatch(null, parquetCachedBatch.numRows).rowIterator().asScala
} else {
val inputFile = new ByteArrayInputFile(parquetCachedBatch.buffer)
new PcbFileToRowsIterator(inputFile, cacheAttributes, hadoopConf, options)
}
}
}
/**
* This iterator wraps an `Iterator[InternalRow]` and will transform the rows so that
* unsupported types are converted back to the original type that was cached.
*/
private class UnsupportedRowTransformIterator(
rowIter: Iterator[InternalRow],
cacheAttributes: Seq[Attribute],
origCacheSchema: Seq[Attribute],
origRequestedSchema: Seq[Attribute]) extends UnsupportedDataHandlerIterator {
val newRow = new GenericInternalRow(cacheAttributes.length)
lazy val unsafeProjection =
GenerateUnsafeProjection.generate(origRequestedSchema, origCacheSchema)
override def hasNext: Boolean = rowIter.hasNext
override def next(): InternalRow = {
//read a row and convert it to what the caller is expecting
val row = rowIter.next()
handleInternalRow(origCacheSchema, row, newRow)
unsafeProjection.apply(newRow)
}
override def handleInterval(
data: SpecializedGetters,
index: Int): CalendarInterval = {
if (data.isNullAt(index)) {
null
} else {
val structData = data.getStruct(index, 3)
new CalendarInterval(structData.getInt(0),
structData.getInt(1), structData.getLong(2))
}
}
override def handleInternalRow(
schema: Seq[Attribute],
row: InternalRow,
newRow: InternalRow): Unit = {
schema.indices.foreach { index =>
val dataType = schema(index).dataType
if (row.isNullAt(index)) {
newRow.setNullAt(index)
} else {
dataType match {
case s@StructType(_) =>
val supportedSchema =
PCBSSchemaHelper.getSupportedDataType(dataType).asInstanceOf[StructType]
val structRow =
handleStruct(row.getStruct(index, supportedSchema.size), s, s)
newRow.update(index, structRow)
case a@ArrayType(_, _) =>
val arrayData = row.getArray(index)
newRow.update(index, handleArray(a.elementType, arrayData))
case MapType(keyType, valueType, _) =>
val mapData = row.getMap(index)
newRow.update(index, handleMap(keyType, valueType, mapData))
case CalendarIntervalType =>
val interval = handleInterval(row, index)
if (interval == null) {
newRow.setNullAt(index)
} else {
newRow.setInterval(index, interval)
}
case d: DecimalType =>
if (row.isNullAt(index)) {
newRow.setDecimal(index, null, d.precision)
} else {
val dec = if (d.precision <= Decimal.MAX_INT_DIGITS) {
Decimal(row.getInt(index).toLong, d.precision, d.scale)
} else {
Decimal(row.getLong(index), d.precision, d.scale)
}
newRow.update(index, dec)
}
case NullType =>
newRow.setNullAt(index)
case _ =>
newRow.update(index, row.get(index, dataType))
}
}
}
}
}
private abstract class UnsupportedDataHandlerIterator extends Iterator[InternalRow] {
def handleInternalRow(schema: Seq[Attribute], row: InternalRow, newRow: InternalRow): Unit
def handleInterval(data: SpecializedGetters, index: Int): Any
def handleStruct(
data: InternalRow,
origSchema: StructType,
supportedSchema: StructType): InternalRow = {
val structRow = InternalRow.fromSeq(supportedSchema)
handleInternalRow(origSchema.map(field =>
AttributeReference(field.name, field.dataType, field.nullable)()), data, structRow)
structRow
}
def handleMap(
keyType: DataType,
valueType: DataType,
mapData: MapData): MapData = {
val keyData = mapData.keyArray()
val newKeyData = handleArray(keyType, keyData)
val valueData = mapData.valueArray()
val newValueData = handleArray(valueType, valueData)
new ArrayBasedMapData(newKeyData, newValueData)
}
def handleArray(
dataType: DataType,
arrayData: ArrayData): ArrayData = {
dataType match {
case s@StructType(_) =>
val listBuffer = new ListBuffer[InternalRow]()
val supportedSchema =
PCBSSchemaHelper.getSupportedDataType(dataType).asInstanceOf[StructType]
arrayData.foreach(supportedSchema, (_, data) => {
val structRow =
handleStruct(data.asInstanceOf[InternalRow], s, s)
listBuffer += structRow.copy()
})
new GenericArrayData(listBuffer)
case ArrayType(elementType, _) =>
val arrayList = new ListBuffer[Any]()
scala.Range(0, arrayData.numElements()).foreach { i =>
val subArrayData = arrayData.getArray(i)
arrayList.append(handleArray(elementType, subArrayData))
}
new GenericArrayData(arrayList)
case m@MapType(_, _, _) =>
val mapList =
new ListBuffer[Any]()
scala.Range(0, arrayData.numElements()).foreach { i =>
val mapData = arrayData.getMap(i)
mapList.append(handleMap(m.keyType, m.valueType, mapData))
}
new GenericArrayData(mapList)
case CalendarIntervalType =>
val citList = new ListBuffer[Any]()
scala.Range(0, arrayData.numElements()).foreach { i =>
val citRow = handleInterval(arrayData, i)
citList += citRow
}
new GenericArrayData(citList)
case _ =>
arrayData
}
}
}
/**
* Consumes the Iterator[CachedBatch] to return either Iterator[ColumnarBatch] or
* Iterator[InternalRow]
*/
private class CachedBatchIteratorConsumer(
cbIter: Iterator[CachedBatch],
cacheAttributes: Seq[Attribute],
selectedAttributes: Seq[Attribute],
origCacheSchema: Seq[Attribute],
origRequestedSchema: Seq[Attribute],
sharedConf: Broadcast[Map[String, String]]) {
private val conf: SQLConf = getConfFromMap(sharedConf)
private val hadoopConf: Configuration = getHadoopConf(origRequestedSchema.toStructType, conf)
private val options: ParquetReadOptions = HadoopReadOptions.builder(hadoopConf).build()
private lazy val hasUnsupportedType = origCacheSchema.exists { field =>
!PCBSSchemaHelper.isTypeSupportedByParquet(field.dataType)
}
private lazy val allSupportedUnsafeProjection =
GenerateUnsafeProjection.generate(selectedAttributes, cacheAttributes)
def getInternalRowIterator: Iterator[InternalRow] = {
val rawRows = PcbToRowsIterator(cbIter, cacheAttributes, hadoopConf,
options)
if (hasUnsupportedType) {
new UnsupportedRowTransformIterator(rawRows, cacheAttributes, origCacheSchema,
origRequestedSchema)
} else {
// Convert the result into unsafe rows, and drop columns we don't want to read
rawRows.map(allSupportedUnsafeProjection)
}
}
/**
* This method returns a ColumnarBatch iterator over a CachedBatch.
* Each CachedBatch => ColumnarBatch is a 1-1 conversion so its pretty straight forward
*/
def getColumnBatchIterator: Iterator[ColumnarBatch] = new Iterator[ColumnarBatch] {
var iter = getIterator
def getIterator: Iterator[ColumnarBatch] = {
if (!cbIter.hasNext) {
Iterator.empty
} else {
val cachedBatch = cbIter.next().asInstanceOf[ParquetCachedBatch]
if (cachedBatch.sizeInBytes == 0) {
// Edge case where we have stored an empty parquet file possibly a dataframe with no
// columns
List(new ColumnarBatch(null, cachedBatch.numRows)).iterator
} else {
new ShimCurrentBatchIterator(
cachedBatch,
conf,
selectedAttributes,
options, hadoopConf)
}
}
}
override def hasNext: Boolean = {
// go over the batch and get the next non-degenerate iterator
// and return if it hasNext
while ((iter == null || !iter.hasNext) && cbIter.hasNext) {
iter = getIterator
}
iter != null && iter.hasNext
}
override def next(): ColumnarBatch = {
// will return the next ColumnarBatch if hasNext() is true, otherwise throw
if (hasNext) {
iter.next()
} else {
throw new NoSuchElementException("no elements found")
}
}
}
}
private def getConfFromMap(sharedConf: Broadcast[Map[String, String]]): SQLConf = {
val conf = new SQLConf()
sharedConf.value.foreach { case (k, v) => conf.setConfString(k, v) }
conf
}
def getBytesAllowedPerBatch(conf: SQLConf): Long = {
val gpuBatchSize = RapidsConf.GPU_BATCH_SIZE_BYTES.get(conf)
// we are rough estimating 0.5% as meta_data_size. we can do better estimation in future
val approxMetaDataSizeBytes = gpuBatchSize * 0.5/100
(gpuBatchSize - approxMetaDataSizeBytes).toLong
}
/**
* This is a private helper class to return Iterator to convert InternalRow or ColumnarBatch to
* CachedBatch. There is no type checking so if the type of T is anything besides InternalRow
* or ColumnarBatch then the behavior is undefined.
*
* @param iter - an iterator over InternalRow or ColumnarBatch
* @param cachedAttributes - Schema of the cached batch
* @param sharedConf - SQL conf
* @tparam T - Strictly either InternalRow or ColumnarBatch
*/
private[rapids] class CachedBatchIteratorProducer[T](
iter: Iterator[T],
cachedAttributes: Seq[Attribute],
origCachedAttributes: Seq[Attribute],
sharedConf: Broadcast[Map[String, String]]) {
val conf: SQLConf = getConfFromMap(sharedConf)
val bytesAllowedPerBatch = getBytesAllowedPerBatch(conf)
val hadoopConf: Configuration = getHadoopConf(cachedAttributes.toStructType, conf)
def getInternalRowToCachedBatchIterator: Iterator[CachedBatch] = {
new InternalRowToCachedBatchIterator
}
def getColumnarBatchToCachedBatchIterator: Iterator[CachedBatch] = {
new ColumnarBatchToCachedBatchIterator
}
/**
* This class produces an Iterator[CachedBatch] from Iterator[InternalRow]. This is a n-1
* relationship. Each partition represents a single parquet file, so we encode it
* and return the CachedBatch when next is called.
*/
class InternalRowToCachedBatchIterator extends Iterator[CachedBatch]() with AutoCloseable {
var parquetOutputFileFormat = new ParquetOutputFileFormat()
// For testing only
private[rapids] def setParquetOutputFileFormat(p: ParquetOutputFileFormat): Unit = {
parquetOutputFileFormat = p
}
// is there a type that spark doesn't support by default in the schema?
val hasUnsupportedType: Boolean = origCachedAttributes.exists { attribute =>
!PCBSSchemaHelper.isTypeSupportedByParquet(attribute.dataType)
}
override def close(): Unit = {
// We don't have any off heap vectors, so nothing to clean
}
def getIterator: Iterator[InternalRow] = {
if (!hasUnsupportedType) {
iter.asInstanceOf[Iterator[InternalRow]]
} else {
new UnsupportedDataHandlerIterator {
val wrappedIter: Iterator[InternalRow] = iter.asInstanceOf[Iterator[InternalRow]]
val newRow: InternalRow = InternalRow.fromSeq(cachedAttributes)
override def hasNext: Boolean = wrappedIter.hasNext
override def next(): InternalRow = {
val row = wrappedIter.next()
handleInternalRow(origCachedAttributes, row, newRow)
newRow
}
override def handleInterval(
data: SpecializedGetters,
index: Int): InternalRow = {
val citRow = InternalRow(IntegerType, IntegerType, LongType)
if (data.isNullAt(index)) {
null
} else {
val cit = data.getInterval(index)
citRow.setInt(0, cit.months)
citRow.setInt(1, cit.days)
citRow.setLong(2, cit.microseconds)
citRow
}
}
override def handleInternalRow(
schema: Seq[Attribute],
row: InternalRow,
newRow: InternalRow): Unit = {
schema.indices.foreach { index =>
val dataType = schema(index).dataType
if (row.isNullAt(index)) {
newRow.setNullAt(index)
} else {
dataType match {
case s@StructType(_) =>
val newSchema =
PCBSSchemaHelper.getSupportedDataType(dataType).asInstanceOf[StructType]
val structRow =
handleStruct(row.getStruct(index, s.fields.length), s, newSchema)
newRow.update(index, structRow)
case ArrayType(arrayDataType, _) =>
val arrayData = row.getArray(index)
val newArrayData = handleArray(arrayDataType, arrayData)
newRow.update(index, newArrayData)
case MapType(keyType, valueType, _) =>
val mapData = row.getMap(index)
val map = handleMap(keyType, valueType, mapData)
newRow.update(index, map)
case CalendarIntervalType =>
val structData: InternalRow = handleInterval(row, index)
if (structData == null) {
newRow.setNullAt(index)
} else {
newRow.update(index, structData)
}
case d: DecimalType if d.scale < 0 =>
if (d.precision <= Decimal.MAX_INT_DIGITS) {
newRow.update(index, row.getDecimal(index, d.precision, d.scale)
.toUnscaledLong.toInt)
} else {
newRow.update(index, row.getDecimal(index, d.precision, d.scale)
.toUnscaledLong)
}
case _ =>
newRow.update(index, row.get(index, dataType))
}
}
}
}
}
}
}
override def hasNext: Boolean = {
if (queue.isEmpty && !iter.hasNext) {
// iterator has exhausted, we should clean up
close()
}
queue.nonEmpty || iter.hasNext
}
private val queue = new mutable.Queue[CachedBatch]()
//estimate the size of a row
val estimatedSize: Int = cachedAttributes.map { attr =>
attr.dataType.defaultSize
}.sum
override def next(): CachedBatch = {
if (queue.isEmpty) {
queue ++= convertRowIteratorToCachedBatch(getIterator)
}
if (!queue.isEmpty) {
queue.dequeue()
} else {
throw new IllegalStateException("Encountered an empty batch")
}
}
protected def convertRowIteratorToCachedBatch(
rowIterator: Iterator[InternalRow]): mutable.Queue[CachedBatch] = {
val queue = new mutable.Queue[CachedBatch]()
// to store a row if we have read it but there is no room in the parquet file to put it
// we will put it in the next CachedBatch
var leftOverRow: Option[InternalRow] = None
while (rowIterator.hasNext || leftOverRow.nonEmpty) {
// Each partition will be a single parquet file
var rows = 0
// at least a single block
val stream = new ByteArrayOutputStream(ByteArrayOutputFile.BLOCK_SIZE)
val outputFile: OutputFile = new ByteArrayOutputFile(stream)
conf.setConfString(SparkShimImpl.parquetRebaseWriteKey,
LegacyBehaviorPolicyShim.CORRECTED_STR)
if (cachedAttributes.isEmpty) {
// The schema is empty, most probably it is because of the edge case where there
// are no columns in the Dataframe but has rows so let's create an empty CachedBatch
// with rows
queue += new ParquetCachedBatch(rowIterator.size, new Array[Byte](0))
return queue
}
val recordWriter = SQLConf.withExistingConf(conf) {
parquetOutputFileFormat.getRecordWriter(outputFile, hadoopConf)
}
var totalSize = 0
while ((rowIterator.hasNext || leftOverRow.nonEmpty)
&& totalSize < bytesAllowedPerBatch) {
val row = if (leftOverRow.nonEmpty) {
val a = leftOverRow.get
leftOverRow = None // reset value
a
} else {
rowIterator.next()
}
totalSize += {
row match {
case r: UnsafeRow =>
r.getSizeInBytes
case _ =>
estimatedSize
}
}
if (totalSize <= bytesAllowedPerBatch) {
rows += 1
if (rows < 0) {
throw new IllegalStateException("CachedBatch doesn't support rows larger " +
"than Int.MaxValue")
}
recordWriter.write(null, row)
} else {
leftOverRow = Some(if (row.isInstanceOf[UnsafeRow]) {
row.copy()
} else {
row
})
}
}
// passing null as context isn't used in this method
recordWriter.close(null)
queue += ParquetCachedBatch(rows, stream.toByteArray)
}
queue
}
}
/**
* This class produces an Iterator[CachedBatch] from Iterator[ColumnarBatch]. This is a 1-1
* relationship. Each ColumnarBatch is converted to a single ParquetCachedBatch when next()
* is called on this iterator
*/
class ColumnarBatchToCachedBatchIterator extends InternalRowToCachedBatchIterator
with AutoCloseable {
val hostBatches = new ListBuffer[ColumnarBatch]()
// Don't install the callback if in a unit test
Option(TaskContext.get()).foreach { tc =>
onTaskCompletion(tc) {
hostBatches.foreach(_.close())
hostBatches.clear()
close()
}
}
override def next(): CachedBatch = myIter.next()
override def hasNext(): Boolean = myIter.hasNext
val myIter = iter.asInstanceOf[Iterator[ColumnarBatch]].flatMap { batch =>
val hostBatch = if (batch.column(0).isInstanceOf[GpuColumnVector]) {
withResource(batch) { batch =>
new ColumnarBatch(batch.safeMap(_.copyToHost()).toArray, batch.numRows())
}
} else {
batch
}
hostBatches += hostBatch
if (batch.numCols() == 0) {
List(ParquetCachedBatch(batch.numRows(), new Array[Byte](0)))
} else {
convertRowIteratorToCachedBatch(hostBatch.rowIterator().asScala).toList
}
}
override def close(): Unit = {
hostBatches.foreach(_.close())
hostBatches.clear()
}
}
}
// We want to change the original schema to have the new names as well
private def sanitizeColumnNames(originalSchema: Seq[Attribute],
schemaToCopyNamesFrom: Seq[Attribute]): Seq[Attribute] = {
originalSchema.zip(schemaToCopyNamesFrom).map {
case (origAttribute, newAttribute) => origAttribute.withName(newAttribute.name)
}
}
private def getSupportedSchemaFromUnsupported(
cachedAttributes: Seq[Attribute],
requestedAttributes: Seq[Attribute] = Seq.empty): (Seq[Attribute], Seq[Attribute]) = {
val newCachedAttributes =
PCBSSchemaHelper.getSupportedSchemaFromUnsupported(cachedAttributes)
val newRequestedAttributes =
getSelectedSchemaFromCachedSchema(requestedAttributes, newCachedAttributes)
(newCachedAttributes, newRequestedAttributes)
}
private def getHadoopConf(requestedSchema: StructType,
sqlConf: SQLConf): Configuration = {
val hadoopConf = new Configuration(false)
hadoopConf.set(ParquetInputFormat.READ_SUPPORT_CLASS, classOf[ParquetReadSupport].getName)
hadoopConf.set(
ParquetReadSupport.SPARK_ROW_REQUESTED_SCHEMA,
requestedSchema.json)
hadoopConf.set(
ParquetWriteSupport.SPARK_ROW_SCHEMA,
requestedSchema.json)
hadoopConf.set(
SQLConf.SESSION_LOCAL_TIMEZONE.key,
sqlConf.sessionLocalTimeZone)
hadoopConf.setBoolean(SQLConf.PARQUET_BINARY_AS_STRING.key, false)
hadoopConf.setBoolean(SQLConf.PARQUET_INT96_AS_TIMESTAMP.key, false)
hadoopConf.setBoolean(SQLConf.CASE_SENSITIVE.key, false)
hadoopConf.set(SparkShimImpl.parquetRebaseWriteKey, LegacyBehaviorPolicyShim.CORRECTED_STR)
hadoopConf.set(SQLConf.PARQUET_OUTPUT_TIMESTAMP_TYPE.key,
SQLConf.ParquetOutputTimestampType.TIMESTAMP_MICROS.toString)
hadoopConf.setBoolean(SQLConf.PARQUET_WRITE_LEGACY_FORMAT.key, false)
hadoopConf.set(ParquetOutputFormat.WRITER_VERSION,
ParquetProperties.WriterVersion.PARQUET_1_0.toString)
ParquetWriteSupport.setSchema(requestedSchema, hadoopConf)
// From 3.3.0, Spark will check this filed ID config
ParquetFieldIdShims.setupParquetFieldIdWriteConfig(hadoopConf, sqlConf)
// timestamp_NTZ flag is introduced in Spark 3.4.0
ParquetTimestampNTZShims.setupTimestampNTZConfig(hadoopConf, sqlConf)
// From 3.3.2, Spark schema converter needs this conf
ParquetLegacyNanoAsLongShims.setupLegacyParquetNanosAsLongForPCBS(hadoopConf)
hadoopConf
}
/**
* Convert an `RDD[InternalRow]` into an `RDD[CachedBatch]` in preparation for caching the data.
* We use the RowToColumnarIterator and convert each batch at a time
*
* @param input the input `RDD` to be converted.
* @param schema the schema of the data being stored.
* @param storageLevel where the data will be stored.
* @param conf the config for the query.
* @return The data converted into a format more suitable for caching.
*/
override def convertInternalRowToCachedBatch(
input: RDD[InternalRow],
schema: Seq[Attribute],
storageLevel: StorageLevel,
conf: SQLConf): RDD[CachedBatch] = {
val rapidsConf = new RapidsConf(conf)
val useCompression = conf.useCompression
val bytesAllowedPerBatch = getBytesAllowedPerBatch(conf)
val (schemaWithUnambiguousNames, _) = getSupportedSchemaFromUnsupported(schema)
if (rapidsConf.isSqlEnabled && rapidsConf.isSqlExecuteOnGPU &&
isSchemaSupportedByCudf(schema)) {
val structSchema = schemaWithUnambiguousNames.toStructType
val converters = new GpuRowToColumnConverter(structSchema)
val columnarBatchRdd = input.mapPartitions(iter => {
new RowToColumnarIterator(iter, structSchema, RequireSingleBatch, converters)
})
columnarBatchRdd.flatMap(cb => {
withResource(cb)(cb => compressColumnarBatchWithParquet(cb, structSchema,
schema.toStructType, bytesAllowedPerBatch, useCompression))
})
} else {
val broadcastedConf = SparkSession.active.sparkContext.broadcast(conf.getAllConfs)
input.mapPartitions {
cbIter =>
new CachedBatchIteratorProducer[InternalRow](cbIter, schemaWithUnambiguousNames, schema,
broadcastedConf).getInternalRowToCachedBatchIterator
}
}
}
override def buildFilter(
predicates: Seq[Expression],
cachedAttributes: Seq[Attribute]): (Int, Iterator[CachedBatch]) => Iterator[CachedBatch] = {
//essentially a noop
(_: Int, b: Iterator[CachedBatch]) => b
}
}
/**
* Similar to ParquetFileFormat
*/
private[rapids] class ParquetOutputFileFormat {
@scala.annotation.nowarn(
"msg=constructor .* in class .* is deprecated"
)
def getRecordWriter(output: OutputFile, conf: Configuration): RecordWriter[Void, InternalRow] = {
import ParquetOutputFormat._
val blockSize = getLongBlockSize(conf)
val maxPaddingSize =
conf.getInt(MAX_PADDING_BYTES, ParquetWriter.MAX_PADDING_SIZE_DEFAULT)
val validating = getValidation(conf)
val writeSupport = new ParquetWriteSupport().asInstanceOf[WriteSupport[InternalRow]]
val init = writeSupport.init(conf)
val writer = new ParquetFileWriter(output, init.getSchema,
Mode.CREATE, blockSize, maxPaddingSize)
writer.start()
val writerVersion =
ParquetProperties.WriterVersion.fromString(conf.get(ParquetOutputFormat.WRITER_VERSION,
ParquetProperties.WriterVersion.PARQUET_1_0.toString))
val codecFactory = new CodecFactory(conf, getPageSize(conf))
new ParquetRecordWriter[InternalRow](writer, writeSupport, init.getSchema,
init.getExtraMetaData, blockSize, getPageSize(conf),
codecFactory.getCompressor(CompressionCodecName.UNCOMPRESSED), getDictionaryPageSize(conf),
getEnableDictionary(conf), validating, writerVersion,
ParquetOutputFileFormat.getMemoryManager(conf))
}
}
private object ParquetOutputFileFormat {
var memoryManager: MemoryManager = _
val DEFAULT_MEMORY_POOL_RATIO: Float = 0.95f
val DEFAULT_MIN_MEMORY_ALLOCATION: Long = 1 * 1024 * 1024 // 1MB
def getMemoryManager(conf: Configuration): MemoryManager = {
synchronized {
if (memoryManager == null) {
import ParquetOutputFormat._
val maxLoad = conf.getFloat(MEMORY_POOL_RATIO, DEFAULT_MEMORY_POOL_RATIO)
val minAllocation = conf.getLong(MIN_MEMORY_ALLOCATION, DEFAULT_MIN_MEMORY_ALLOCATION)
memoryManager = new MemoryManager(maxLoad, minAllocation)
}
}
memoryManager
}
}