[GitHub] [spark] HeartSaVioR commented on a diff in pull request #37893: [DRAFT][DO-NOT-MERGE][SPARK-40434][SS][PYTHON] Implement applyInPandasWithState in PySpark

[GitBox <gi...@apache.org>]

HeartSaVioR commented on code in PR #37893:
URL: https://github.com/apache/spark/pull/37893#discussion_r971613565


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/python/FlatMapGroupsInPandasWithStateExec.scala:
##########
@@ -0,0 +1,217 @@
+/*
+ * 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.execution.python
+
+import org.apache.spark.TaskContext
+import org.apache.spark.api.python.{ChainedPythonFunctions, PythonEvalType}
+import org.apache.spark.sql.Row
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.encoders.{ExpressionEncoder, RowEncoder}
+import org.apache.spark.sql.catalyst.expressions._
+import org.apache.spark.sql.catalyst.plans.logical.{EventTimeTimeout, ProcessingTimeTimeout}
+import org.apache.spark.sql.catalyst.plans.physical.Distribution
+import org.apache.spark.sql.execution.{GroupedIterator, SparkPlan, UnaryExecNode}
+import org.apache.spark.sql.execution.python.PandasGroupUtils.resolveArgOffsets
+import org.apache.spark.sql.execution.streaming._
+import org.apache.spark.sql.execution.streaming.GroupStateImpl.NO_TIMESTAMP
+import org.apache.spark.sql.execution.streaming.state.FlatMapGroupsWithStateExecHelper.StateData
+import org.apache.spark.sql.execution.streaming.state.StateStore
+import org.apache.spark.sql.streaming.{GroupStateTimeout, OutputMode}
+import org.apache.spark.sql.types.StructType
+import org.apache.spark.sql.util.ArrowUtils
+import org.apache.spark.util.CompletionIterator
+
+/**
+ * Physical operator for executing
+ * [[org.apache.spark.sql.catalyst.plans.logical.FlatMapGroupsInPandasWithState]]
+ *
+ * @param functionExpr function called on each group
+ * @param groupingAttributes used to group the data
+ * @param outAttributes used to define the output rows
+ * @param stateType used to serialize/deserialize state before calling `functionExpr`
+ * @param stateInfo `StatefulOperatorStateInfo` to identify the state store for a given operator.
+ * @param stateFormatVersion the version of state format.
+ * @param outputMode the output mode of `functionExpr`
+ * @param timeoutConf used to timeout groups that have not received data in a while
+ * @param batchTimestampMs processing timestamp of the current batch.
+ * @param eventTimeWatermark event time watermark for the current batch
+ * @param child logical plan of the underlying data
+ */
+case class FlatMapGroupsInPandasWithStateExec(
+    functionExpr: Expression,
+    groupingAttributes: Seq[Attribute],
+    outAttributes: Seq[Attribute],
+    stateType: StructType,
+    stateInfo: Option[StatefulOperatorStateInfo],
+    stateFormatVersion: Int,
+    outputMode: OutputMode,
+    timeoutConf: GroupStateTimeout,
+    batchTimestampMs: Option[Long],
+    eventTimeWatermark: Option[Long],
+    child: SparkPlan) extends UnaryExecNode with FlatMapGroupsWithStateExecBase {
+
+  // TODO(SPARK-XXXXX): Add the support of initial state.

Review Comment:
   self-comment: It does not seem to be one of first priority functionalities - I think we can deal with this later. I'll file a JIRA ticket and change the ticket number. 



##########
sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala:
##########
@@ -793,6 +812,9 @@ abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
           initialStateGroupAttrs, data, initialStateDataAttrs, output, timeout,
           hasInitialState, planLater(initialState), planLater(child)
         ) :: Nil
+      case _: FlatMapGroupsInPandasWithState =>
+        throw new UnsupportedOperationException(

Review Comment:
   In Scala/Java, flatMapGroupsWithState API uses different physical plan for batch query, which applies some wrap-up with user function to effectively match up with user function in MapGroupsExec.
   
   This seems to be non-trivial to deal with in PySpark, or at least seems to require major work. As there is another API for batch-friendly query (applyInPandas), I think we can leave this to TODO and guide users to use applyInPandas instead for now.
   
   self-comment: file a JIRA ticket and replace the number.



-- 
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