You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by GitBox <gi...@apache.org> on 2022/02/08 23:04:15 UTC
[GitHub] [spark] xuanyuanking commented on a change in pull request #35259: [SPARK-37970][SS] Introduce AcceptsLatestSeenOffset to indicate latest seen offset to streaming source
xuanyuanking commented on a change in pull request #35259:
URL: https://github.com/apache/spark/pull/35259#discussion_r802131955
##########
File path: sql/core/src/test/scala/org/apache/spark/sql/streaming/AcceptsLatestSeenOffsetSuite.scala
##########
@@ -0,0 +1,272 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You 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 org.apache.spark.sql.streaming
+
+import org.scalatest.BeforeAndAfter
+
+import org.apache.spark.sql._
+import org.apache.spark.sql.catalyst.plans.logical.Range
+import org.apache.spark.sql.connector.read.streaming
+import org.apache.spark.sql.connector.read.streaming.{AcceptsLatestSeenOffset, SparkDataStream}
+import org.apache.spark.sql.execution.streaming._
+import org.apache.spark.sql.execution.streaming.sources.{ContinuousMemoryStream, ContinuousMemoryStreamOffset}
+import org.apache.spark.sql.types.{LongType, StructType}
+
+class AcceptsLatestSeenOffsetSuite extends StreamTest with BeforeAndAfter {
+
+ import testImplicits._
+
+ after {
+ sqlContext.streams.active.foreach(_.stop())
+ }
+
+ test("DataSource V1 source with micro-batch is not supported") {
+ val testSource = new TestSource(spark)
+ val df = testSource.toDF()
+
+ /** Add data to this test source by incrementing its available offset */
+ def addData(numNewRows: Int): StreamAction = new AddData {
+ override def addData(
+ query: Option[StreamExecution]): (SparkDataStream, streaming.Offset) = {
+ testSource.incrementAvailableOffset(numNewRows)
+ (testSource, testSource.getOffset.get)
+ }
+ }
+
+ addData(10)
+ val query = df.writeStream.format("console").start()
+ val exc = intercept[StreamingQueryException] {
+ query.processAllAvailable()
+ }
+ assert(exc.getMessage.contains(
+ "AcceptsLatestSeenOffset is not supported with DSv1 streaming source"))
+ }
+
+ test("DataSource V2 source with micro-batch") {
+ val inputData = new TestMemoryStream[Long](0, spark.sqlContext)
+ val df = inputData.toDF().select("value")
+
+ /** Add data to this test source by incrementing its available offset */
+ def addData(values: Array[Long]): StreamAction = new AddData {
+ override def addData(
+ query: Option[StreamExecution]): (SparkDataStream, streaming.Offset) = {
+ (inputData, inputData.addData(values))
+ }
+ }
+
+ testStream(df)(
+ StartStream(),
+ addData((1L to 10L).toArray),
+ ProcessAllAvailable(),
+ Execute("latest seen offset should be null") { _ =>
+ // this verifies that the callback method is not called for the new query
+ assert(inputData.latestSeenOffset === null)
+ },
+ StopStream,
+
+ StartStream(),
+ addData((11L to 20L).toArray),
+ ProcessAllAvailable(),
+ Execute("latest seen offset should be 0") { _ =>
+ assert(inputData.latestSeenOffset === LongOffset(0))
+ },
+ StopStream,
+
+ Execute("mark last batch as incomplete") { q =>
+ // Delete the last committed batch from the commit log to signify that the last batch
+ // (a no-data batch) did not complete and has to be re-executed on restart.
+ val commit = q.commitLog.getLatest().map(_._1).getOrElse(-1L)
+ q.commitLog.purgeAfter(commit - 1)
+ },
+ StartStream(),
+ addData((21L to 30L).toArray),
+ ProcessAllAvailable(),
+ Execute("latest seen offset should be 1") { _ =>
+ assert(inputData.latestSeenOffset === LongOffset(1))
+ }
+ )
+ }
+
+ test("DataSource V2 source with micro-batch - rollback of microbatch 0") {
+ // Test case: when the query is restarted, we expect the execution to call `latestSeenOffset`
+ // first. Later as part of the execution, execution may call `initialOffset` if the previous
+ // run of the query had no committed batches.
+ val inputData = new TestMemoryStream[Long](0, spark.sqlContext)
+ val df = inputData.toDF().select("value")
+
+ /** Add data to this test source by incrementing its available offset */
+ def addData(values: Array[Long]): StreamAction = new AddData {
+ override def addData(
+ query: Option[StreamExecution]): (SparkDataStream, streaming.Offset) = {
+ (inputData, inputData.addData(values))
+ }
+ }
+
+ testStream(df)(
+ StartStream(),
+ addData((1L to 10L).toArray),
+ ProcessAllAvailable(),
+ Execute("latest seen offset should be null") { _ =>
+ // this verifies that the callback method is not called for the new query
+ assert(inputData.latestSeenOffset === null)
+ },
+ StopStream,
+
+ Execute("mark last batch as incomplete") { q =>
+ // Delete the last committed batch from the commit log to signify that the last batch
+ // (a no-data batch) did not complete and has to be re-executed on restart.
+ val commit = q.commitLog.getLatest().map(_._1).getOrElse(-1L)
+ q.commitLog.purgeAfter(commit - 1)
+ },
+
+ Execute("reset flag initial offset called flag") { q =>
+ inputData.assertInitialOffsetIsCalledAfterLatestOffsetSeen = true
Review comment:
Following the description of this test scenario, why we don't set this in `setLatestSeenOffset`?
##########
File path: sql/core/src/test/scala/org/apache/spark/sql/streaming/AcceptsLatestSeenOffsetSuite.scala
##########
@@ -0,0 +1,272 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You 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 org.apache.spark.sql.streaming
+
+import org.scalatest.BeforeAndAfter
+
+import org.apache.spark.sql._
+import org.apache.spark.sql.catalyst.plans.logical.Range
+import org.apache.spark.sql.connector.read.streaming
+import org.apache.spark.sql.connector.read.streaming.{AcceptsLatestSeenOffset, SparkDataStream}
+import org.apache.spark.sql.execution.streaming._
+import org.apache.spark.sql.execution.streaming.sources.{ContinuousMemoryStream, ContinuousMemoryStreamOffset}
+import org.apache.spark.sql.types.{LongType, StructType}
+
+class AcceptsLatestSeenOffsetSuite extends StreamTest with BeforeAndAfter {
+
+ import testImplicits._
+
+ after {
+ sqlContext.streams.active.foreach(_.stop())
+ }
+
+ test("DataSource V1 source with micro-batch is not supported") {
+ val testSource = new TestSource(spark)
+ val df = testSource.toDF()
+
+ /** Add data to this test source by incrementing its available offset */
+ def addData(numNewRows: Int): StreamAction = new AddData {
+ override def addData(
+ query: Option[StreamExecution]): (SparkDataStream, streaming.Offset) = {
+ testSource.incrementAvailableOffset(numNewRows)
+ (testSource, testSource.getOffset.get)
+ }
+ }
+
+ addData(10)
+ val query = df.writeStream.format("console").start()
+ val exc = intercept[StreamingQueryException] {
+ query.processAllAvailable()
+ }
+ assert(exc.getMessage.contains(
+ "AcceptsLatestSeenOffset is not supported with DSv1 streaming source"))
+ }
+
+ test("DataSource V2 source with micro-batch") {
+ val inputData = new TestMemoryStream[Long](0, spark.sqlContext)
+ val df = inputData.toDF().select("value")
+
+ /** Add data to this test source by incrementing its available offset */
+ def addData(values: Array[Long]): StreamAction = new AddData {
+ override def addData(
+ query: Option[StreamExecution]): (SparkDataStream, streaming.Offset) = {
+ (inputData, inputData.addData(values))
+ }
+ }
+
+ testStream(df)(
+ StartStream(),
+ addData((1L to 10L).toArray),
+ ProcessAllAvailable(),
+ Execute("latest seen offset should be null") { _ =>
+ // this verifies that the callback method is not called for the new query
+ assert(inputData.latestSeenOffset === null)
+ },
+ StopStream,
+
+ StartStream(),
+ addData((11L to 20L).toArray),
+ ProcessAllAvailable(),
+ Execute("latest seen offset should be 0") { _ =>
+ assert(inputData.latestSeenOffset === LongOffset(0))
+ },
+ StopStream,
+
+ Execute("mark last batch as incomplete") { q =>
+ // Delete the last committed batch from the commit log to signify that the last batch
+ // (a no-data batch) did not complete and has to be re-executed on restart.
+ val commit = q.commitLog.getLatest().map(_._1).getOrElse(-1L)
+ q.commitLog.purgeAfter(commit - 1)
+ },
+ StartStream(),
+ addData((21L to 30L).toArray),
+ ProcessAllAvailable(),
+ Execute("latest seen offset should be 1") { _ =>
+ assert(inputData.latestSeenOffset === LongOffset(1))
+ }
+ )
+ }
+
+ test("DataSource V2 source with micro-batch - rollback of microbatch 0") {
+ // Test case: when the query is restarted, we expect the execution to call `latestSeenOffset`
+ // first. Later as part of the execution, execution may call `initialOffset` if the previous
+ // run of the query had no committed batches.
+ val inputData = new TestMemoryStream[Long](0, spark.sqlContext)
+ val df = inputData.toDF().select("value")
+
+ /** Add data to this test source by incrementing its available offset */
+ def addData(values: Array[Long]): StreamAction = new AddData {
+ override def addData(
+ query: Option[StreamExecution]): (SparkDataStream, streaming.Offset) = {
+ (inputData, inputData.addData(values))
+ }
+ }
+
+ testStream(df)(
+ StartStream(),
+ addData((1L to 10L).toArray),
+ ProcessAllAvailable(),
+ Execute("latest seen offset should be null") { _ =>
+ // this verifies that the callback method is not called for the new query
+ assert(inputData.latestSeenOffset === null)
+ },
+ StopStream,
+
+ Execute("mark last batch as incomplete") { q =>
+ // Delete the last committed batch from the commit log to signify that the last batch
+ // (a no-data batch) did not complete and has to be re-executed on restart.
+ val commit = q.commitLog.getLatest().map(_._1).getOrElse(-1L)
+ q.commitLog.purgeAfter(commit - 1)
+ },
+
+ Execute("reset flag initial offset called flag") { q =>
+ inputData.assertInitialOffsetIsCalledAfterLatestOffsetSeen = true
+ },
+ StartStream(),
+ addData((11L to 20L).toArray),
+ ProcessAllAvailable(),
+ Execute("latest seen offset should be 0") { _ =>
+ assert(inputData.latestSeenOffset === LongOffset(0))
+ },
+ StopStream
+ )
+ }
+
+ test("DataSource V2 source with continuous mode") {
+ val inputData = new TestContinuousMemoryStream[Long](0, spark.sqlContext, 1)
+ val df = inputData.toDF().select("value")
+
+ /** Add data to this test source by incrementing its available offset */
+ def addData(values: Array[Long]): StreamAction = new AddData {
+ override def addData(
+ query: Option[StreamExecution]): (SparkDataStream, streaming.Offset) = {
+ (inputData, inputData.addData(values))
+ }
+ }
+
+ testStream(df)(
+ StartStream(trigger = Trigger.Continuous("1 hour")),
+ addData((1L to 10L).toArray),
+ AwaitEpoch(0),
+ Execute { _ =>
+ assert(inputData.latestSeenOffset === null)
+ },
+ IncrementEpoch(),
+ StopStream,
+
+ StartStream(trigger = Trigger.Continuous("1 hour")),
+ addData((11L to 20L).toArray),
+ AwaitEpoch(2),
+ Execute { _ =>
+ assert(inputData.latestSeenOffset === ContinuousMemoryStreamOffset(Map(0 -> 10)))
+ },
+ IncrementEpoch(),
+ StopStream,
+
+ StartStream(trigger = Trigger.Continuous("1 hour")),
+ addData((21L to 30L).toArray),
+ AwaitEpoch(3),
+ Execute { _ =>
+ assert(inputData.latestSeenOffset === ContinuousMemoryStreamOffset(Map(0 -> 20)))
+ }
+ )
+ }
+
+ class TestSource(spark: SparkSession) extends Source with AcceptsLatestSeenOffset {
+
+ @volatile var currentOffset = 0L
+
+ override def getOffset: Option[Offset] = {
+ if (currentOffset <= 0) None else Some(LongOffset(currentOffset))
+ }
+
+ override def getBatch(start: Option[Offset], end: Offset): DataFrame = {
+ if (currentOffset == 0) currentOffset = getOffsetValue(end)
+ val plan = Range(
+ start.map(getOffsetValue).getOrElse(0L) + 1L, getOffsetValue(end) + 1L, 1, None,
+ isStreaming = true)
+ Dataset.ofRows(spark, plan)
+ }
+
+ def incrementAvailableOffset(numNewRows: Int): Unit = {
+ currentOffset = currentOffset + numNewRows
+ }
+
+ override def setLatestSeenOffset(offset: streaming.Offset): Unit = {
+ assert(false, "This method should not be called!")
+ }
+
+ def reset(): Unit = {
+ currentOffset = 0L
+ }
+
+ def toDF(): DataFrame = Dataset.ofRows(spark, StreamingExecutionRelation(this, spark))
+ override def schema: StructType = new StructType().add("value", LongType)
+ override def stop(): Unit = {}
+ private def getOffsetValue(offset: Offset): Long = {
+ offset match {
+ case s: SerializedOffset => LongOffset(s).offset
+ case l: LongOffset => l.offset
+ case _ => throw new IllegalArgumentException("incorrect offset type: " + offset)
+ }
+ }
+ }
+
+ class TestMemoryStream[A : Encoder](
+ _id: Int,
+ _sqlContext: SQLContext,
+ _numPartitions: Option[Int] = None)
+ extends MemoryStream[A](_id, _sqlContext, _numPartitions)
+ with AcceptsLatestSeenOffset {
+
+ @volatile var latestSeenOffset: streaming.Offset = null
+
+ // Flag to assert the sequence of calls in following scenario:
+ // When the query is restarted, we expect the execution to call `latestSeenOffset` first.
+ // Later as part of the execution, execution may call `initialOffset` if the previous
+ // run of the query had no committed batches.
+ @volatile var assertInitialOffsetIsCalledAfterLatestOffsetSeen: Boolean = false
+
+ override def setLatestSeenOffset(offset: streaming.Offset): Unit = {
+ latestSeenOffset = offset
+ }
+
+ override def initialOffset: streaming.Offset = {
+ if (assertInitialOffsetIsCalledAfterLatestOffsetSeen) {
Review comment:
nit: if (assertInitialOffsetIsCalledAfterLatestOffsetSeen && latestSeenOffset == null) ?
##########
File path: sql/catalyst/src/main/java/org/apache/spark/sql/connector/read/streaming/AcceptsLatestSeenOffset.java
##########
@@ -0,0 +1,37 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You 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 org.apache.spark.sql.connector.read.streaming;
+
+/**
+ * Indicates that the source accepts latest seen offset, which requires streaming execution
Review comment:
nit: the latest seen offset
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org