From 8d05a7a98bdbd3ce7c81d273e05a375877ebe68f Mon Sep 17 00:00:00 2001 From: hyukjinkwon Date: Tue, 17 May 2016 11:18:51 -0700 Subject: [PATCH] [SPARK-10216][SQL] Avoid creating empty files during overwriting with group by query ## What changes were proposed in this pull request? Currently, `INSERT INTO` with `GROUP BY` query tries to make at least 200 files (default value of `spark.sql.shuffle.partition`), which results in lots of empty files. This PR makes it avoid creating empty files during overwriting into Hive table and in internal data sources with group by query. This checks whether the given partition has data in it or not and creates/writes file only when it actually has data. ## How was this patch tested? Unittests in `InsertIntoHiveTableSuite` and `HadoopFsRelationTest`. Closes #8411 Author: hyukjinkwon Author: Keuntae Park Closes #12855 from HyukjinKwon/pr/8411. --- .../datasources/WriterContainer.scala | 221 +++++++++--------- .../spark/sql/hive/hiveWriterContainers.scala | 24 +- .../sql/hive/InsertIntoHiveTableSuite.scala | 41 +++- .../sql/sources/HadoopFsRelationTest.scala | 22 +- 4 files changed, 182 insertions(+), 126 deletions(-) diff --git a/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/WriterContainer.scala b/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/WriterContainer.scala index 3b064a5bc489f..7e12bbb2128bf 100644 --- a/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/WriterContainer.scala +++ b/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/WriterContainer.scala @@ -239,48 +239,50 @@ private[sql] class DefaultWriterContainer( extends BaseWriterContainer(relation, job, isAppend) { def writeRows(taskContext: TaskContext, iterator: Iterator[InternalRow]): Unit = { - executorSideSetup(taskContext) - val configuration = taskAttemptContext.getConfiguration - configuration.set("spark.sql.sources.output.path", outputPath) - var writer = newOutputWriter(getWorkPath) - writer.initConverter(dataSchema) - - // If anything below fails, we should abort the task. - try { - Utils.tryWithSafeFinallyAndFailureCallbacks { - while (iterator.hasNext) { - val internalRow = iterator.next() - writer.writeInternal(internalRow) - } - commitTask() - }(catchBlock = abortTask()) - } catch { - case t: Throwable => - throw new SparkException("Task failed while writing rows", t) - } + if (iterator.hasNext) { + executorSideSetup(taskContext) + val configuration = taskAttemptContext.getConfiguration + configuration.set("spark.sql.sources.output.path", outputPath) + var writer = newOutputWriter(getWorkPath) + writer.initConverter(dataSchema) - def commitTask(): Unit = { + // If anything below fails, we should abort the task. try { - if (writer != null) { - writer.close() - writer = null - } - super.commitTask() + Utils.tryWithSafeFinallyAndFailureCallbacks { + while (iterator.hasNext) { + val internalRow = iterator.next() + writer.writeInternal(internalRow) + } + commitTask() + }(catchBlock = abortTask()) } catch { - case cause: Throwable => - // This exception will be handled in `InsertIntoHadoopFsRelation.insert$writeRows`, and - // will cause `abortTask()` to be invoked. - throw new RuntimeException("Failed to commit task", cause) + case t: Throwable => + throw new SparkException("Task failed while writing rows", t) } - } - def abortTask(): Unit = { - try { - if (writer != null) { - writer.close() + def commitTask(): Unit = { + try { + if (writer != null) { + writer.close() + writer = null + } + super.commitTask() + } catch { + case cause: Throwable => + // This exception will be handled in `InsertIntoHadoopFsRelation.insert$writeRows`, and + // will cause `abortTask()` to be invoked. + throw new RuntimeException("Failed to commit task", cause) + } + } + + def abortTask(): Unit = { + try { + if (writer != null) { + writer.close() + } + } finally { + super.abortTask() } - } finally { - super.abortTask() } } } @@ -363,84 +365,87 @@ private[sql] class DynamicPartitionWriterContainer( } def writeRows(taskContext: TaskContext, iterator: Iterator[InternalRow]): Unit = { - executorSideSetup(taskContext) - - // We should first sort by partition columns, then bucket id, and finally sorting columns. - val sortingExpressions: Seq[Expression] = partitionColumns ++ bucketIdExpression ++ sortColumns - val getSortingKey = UnsafeProjection.create(sortingExpressions, inputSchema) - - val sortingKeySchema = StructType(sortingExpressions.map { - case a: Attribute => StructField(a.name, a.dataType, a.nullable) - // The sorting expressions are all `Attribute` except bucket id. - case _ => StructField("bucketId", IntegerType, nullable = false) - }) - - // Returns the data columns to be written given an input row - val getOutputRow = UnsafeProjection.create(dataColumns, inputSchema) - - // Returns the partition path given a partition key. - val getPartitionString = - UnsafeProjection.create(Concat(partitionStringExpression) :: Nil, partitionColumns) - - // Sorts the data before write, so that we only need one writer at the same time. - // TODO: inject a local sort operator in planning. - val sorter = new UnsafeKVExternalSorter( - sortingKeySchema, - StructType.fromAttributes(dataColumns), - SparkEnv.get.blockManager, - SparkEnv.get.serializerManager, - TaskContext.get().taskMemoryManager().pageSizeBytes) - - while (iterator.hasNext) { - val currentRow = iterator.next() - sorter.insertKV(getSortingKey(currentRow), getOutputRow(currentRow)) - } - logInfo(s"Sorting complete. Writing out partition files one at a time.") - - val getBucketingKey: InternalRow => InternalRow = if (sortColumns.isEmpty) { - identity - } else { - UnsafeProjection.create(sortingExpressions.dropRight(sortColumns.length).zipWithIndex.map { - case (expr, ordinal) => BoundReference(ordinal, expr.dataType, expr.nullable) + if (iterator.hasNext) { + executorSideSetup(taskContext) + + // We should first sort by partition columns, then bucket id, and finally sorting columns. + val sortingExpressions: Seq[Expression] = + partitionColumns ++ bucketIdExpression ++ sortColumns + val getSortingKey = UnsafeProjection.create(sortingExpressions, inputSchema) + + val sortingKeySchema = StructType(sortingExpressions.map { + case a: Attribute => StructField(a.name, a.dataType, a.nullable) + // The sorting expressions are all `Attribute` except bucket id. + case _ => StructField("bucketId", IntegerType, nullable = false) }) - } - val sortedIterator = sorter.sortedIterator() + // Returns the data columns to be written given an input row + val getOutputRow = UnsafeProjection.create(dataColumns, inputSchema) + + // Returns the partition path given a partition key. + val getPartitionString = + UnsafeProjection.create(Concat(partitionStringExpression) :: Nil, partitionColumns) + + // Sorts the data before write, so that we only need one writer at the same time. + // TODO: inject a local sort operator in planning. + val sorter = new UnsafeKVExternalSorter( + sortingKeySchema, + StructType.fromAttributes(dataColumns), + SparkEnv.get.blockManager, + SparkEnv.get.serializerManager, + TaskContext.get().taskMemoryManager().pageSizeBytes) + + while (iterator.hasNext) { + val currentRow = iterator.next() + sorter.insertKV(getSortingKey(currentRow), getOutputRow(currentRow)) + } + logInfo(s"Sorting complete. Writing out partition files one at a time.") + + val getBucketingKey: InternalRow => InternalRow = if (sortColumns.isEmpty) { + identity + } else { + UnsafeProjection.create(sortingExpressions.dropRight(sortColumns.length).zipWithIndex.map { + case (expr, ordinal) => BoundReference(ordinal, expr.dataType, expr.nullable) + }) + } - // If anything below fails, we should abort the task. - var currentWriter: OutputWriter = null - try { - Utils.tryWithSafeFinallyAndFailureCallbacks { - var currentKey: UnsafeRow = null - while (sortedIterator.next()) { - val nextKey = getBucketingKey(sortedIterator.getKey).asInstanceOf[UnsafeRow] - if (currentKey != nextKey) { - if (currentWriter != null) { - currentWriter.close() - currentWriter = null - } - currentKey = nextKey.copy() - logDebug(s"Writing partition: $currentKey") + val sortedIterator = sorter.sortedIterator() - currentWriter = newOutputWriter(currentKey, getPartitionString) + // If anything below fails, we should abort the task. + var currentWriter: OutputWriter = null + try { + Utils.tryWithSafeFinallyAndFailureCallbacks { + var currentKey: UnsafeRow = null + while (sortedIterator.next()) { + val nextKey = getBucketingKey(sortedIterator.getKey).asInstanceOf[UnsafeRow] + if (currentKey != nextKey) { + if (currentWriter != null) { + currentWriter.close() + currentWriter = null + } + currentKey = nextKey.copy() + logDebug(s"Writing partition: $currentKey") + + currentWriter = newOutputWriter(currentKey, getPartitionString) + } + currentWriter.writeInternal(sortedIterator.getValue) + } + if (currentWriter != null) { + currentWriter.close() + currentWriter = null } - currentWriter.writeInternal(sortedIterator.getValue) - } - if (currentWriter != null) { - currentWriter.close() - currentWriter = null - } - commitTask() - }(catchBlock = { - if (currentWriter != null) { - currentWriter.close() - } - abortTask() - }) - } catch { - case t: Throwable => - throw new SparkException("Task failed while writing rows", t) + commitTask() + }(catchBlock = { + if (currentWriter != null) { + currentWriter.close() + } + abortTask() + }) + } catch { + case t: Throwable => + throw new SparkException("Task failed while writing rows", t) + } } } } diff --git a/sql/hive/src/main/scala/org/apache/spark/sql/hive/hiveWriterContainers.scala b/sql/hive/src/main/scala/org/apache/spark/sql/hive/hiveWriterContainers.scala index 794fe264ead5d..706fdbc2604fe 100644 --- a/sql/hive/src/main/scala/org/apache/spark/sql/hive/hiveWriterContainers.scala +++ b/sql/hive/src/main/scala/org/apache/spark/sql/hive/hiveWriterContainers.scala @@ -178,19 +178,21 @@ private[hive] class SparkHiveWriterContainer( // this function is executed on executor side def writeToFile(context: TaskContext, iterator: Iterator[InternalRow]): Unit = { - val (serializer, standardOI, fieldOIs, dataTypes, wrappers, outputData) = prepareForWrite() - executorSideSetup(context.stageId, context.partitionId, context.attemptNumber) - - iterator.foreach { row => - var i = 0 - while (i < fieldOIs.length) { - outputData(i) = if (row.isNullAt(i)) null else wrappers(i)(row.get(i, dataTypes(i))) - i += 1 + if (iterator.hasNext) { + val (serializer, standardOI, fieldOIs, dataTypes, wrappers, outputData) = prepareForWrite() + executorSideSetup(context.stageId, context.partitionId, context.attemptNumber) + + iterator.foreach { row => + var i = 0 + while (i < fieldOIs.length) { + outputData(i) = if (row.isNullAt(i)) null else wrappers(i)(row.get(i, dataTypes(i))) + i += 1 + } + writer.write(serializer.serialize(outputData, standardOI)) } - writer.write(serializer.serialize(outputData, standardOI)) - } - close() + close() + } } } diff --git a/sql/hive/src/test/scala/org/apache/spark/sql/hive/InsertIntoHiveTableSuite.scala b/sql/hive/src/test/scala/org/apache/spark/sql/hive/InsertIntoHiveTableSuite.scala index 82d3e49f929d0..883cdac110e0b 100644 --- a/sql/hive/src/test/scala/org/apache/spark/sql/hive/InsertIntoHiveTableSuite.scala +++ b/sql/hive/src/test/scala/org/apache/spark/sql/hive/InsertIntoHiveTableSuite.scala @@ -19,13 +19,13 @@ package org.apache.spark.sql.hive import java.io.File -import org.apache.hadoop.hive.conf.HiveConf import org.scalatest.BeforeAndAfter import org.apache.spark.SparkException -import org.apache.spark.sql.{QueryTest, _} +import org.apache.spark.sql._ import org.apache.spark.sql.catalyst.plans.logical.InsertIntoTable import org.apache.spark.sql.hive.test.TestHiveSingleton +import org.apache.spark.sql.internal.SQLConf import org.apache.spark.sql.test.SQLTestUtils import org.apache.spark.sql.types._ import org.apache.spark.util.Utils @@ -118,10 +118,10 @@ class InsertIntoHiveTableSuite extends QueryTest with TestHiveSingleton with Bef sql( s""" - |CREATE TABLE table_with_partition(c1 string) - |PARTITIONED by (p1 string,p2 string,p3 string,p4 string,p5 string) - |location '${tmpDir.toURI.toString}' - """.stripMargin) + |CREATE TABLE table_with_partition(c1 string) + |PARTITIONED by (p1 string,p2 string,p3 string,p4 string,p5 string) + |location '${tmpDir.toURI.toString}' + """.stripMargin) sql( """ |INSERT OVERWRITE TABLE table_with_partition @@ -216,6 +216,35 @@ class InsertIntoHiveTableSuite extends QueryTest with TestHiveSingleton with Bef sql("DROP TABLE hiveTableWithStructValue") } + test("SPARK-10216: Avoid empty files during overwrite into Hive table with group by query") { + withSQLConf(SQLConf.SHUFFLE_PARTITIONS.key -> "10") { + val testDataset = hiveContext.sparkContext.parallelize( + (1 to 2).map(i => TestData(i, i.toString))).toDF() + testDataset.createOrReplaceTempView("testDataset") + + val tmpDir = Utils.createTempDir() + sql( + s""" + |CREATE TABLE table1(key int,value string) + |location '${tmpDir.toURI.toString}' + """.stripMargin) + sql( + """ + |INSERT OVERWRITE TABLE table1 + |SELECT count(key), value FROM testDataset GROUP BY value + """.stripMargin) + + val overwrittenFiles = tmpDir.listFiles() + .filter(f => f.isFile && !f.getName.endsWith(".crc")) + .sortBy(_.getName) + val overwrittenFilesWithoutEmpty = overwrittenFiles.filter(_.length > 0) + + assert(overwrittenFiles === overwrittenFilesWithoutEmpty) + + sql("DROP TABLE table1") + } + } + test("Reject partitioning that does not match table") { withSQLConf(("hive.exec.dynamic.partition.mode", "nonstrict")) { sql("CREATE TABLE partitioned (id bigint, data string) PARTITIONED BY (part string)") diff --git a/sql/hive/src/test/scala/org/apache/spark/sql/sources/HadoopFsRelationTest.scala b/sql/hive/src/test/scala/org/apache/spark/sql/sources/HadoopFsRelationTest.scala index f4d63334b6573..78d2dc28d6b5e 100644 --- a/sql/hive/src/test/scala/org/apache/spark/sql/sources/HadoopFsRelationTest.scala +++ b/sql/hive/src/test/scala/org/apache/spark/sql/sources/HadoopFsRelationTest.scala @@ -29,7 +29,7 @@ import org.apache.parquet.hadoop.ParquetOutputCommitter import org.apache.spark.deploy.SparkHadoopUtil import org.apache.spark.sql._ import org.apache.spark.sql.execution.DataSourceScanExec -import org.apache.spark.sql.execution.datasources.{FileScanRDD, HadoopFsRelation, LocalityTestFileSystem, LogicalRelation} +import org.apache.spark.sql.execution.datasources.{FileScanRDD, LocalityTestFileSystem} import org.apache.spark.sql.hive.test.TestHiveSingleton import org.apache.spark.sql.internal.SQLConf import org.apache.spark.sql.test.SQLTestUtils @@ -879,6 +879,26 @@ abstract class HadoopFsRelationTest extends QueryTest with SQLTestUtils with Tes } } } + + test("SPARK-10216: Avoid empty files during overwriting with group by query") { + withSQLConf(SQLConf.SHUFFLE_PARTITIONS.key -> "10") { + withTempPath { path => + val df = spark.range(0, 5) + val groupedDF = df.groupBy("id").count() + groupedDF.write + .format(dataSourceName) + .mode(SaveMode.Overwrite) + .save(path.getCanonicalPath) + + val overwrittenFiles = path.listFiles() + .filter(f => f.isFile && !f.getName.startsWith(".") && !f.getName.startsWith("_")) + .sortBy(_.getName) + val overwrittenFilesWithoutEmpty = overwrittenFiles.filter(_.length > 0) + + assert(overwrittenFiles === overwrittenFilesWithoutEmpty) + } + } + } } // This class is used to test SPARK-8578. We should not use any custom output committer when