brkyvz commented on a change in pull request #28040: [SPARK-31278][SS] Fix 
StreamingQuery output rows metric
URL: https://github.com/apache/spark/pull/28040#discussion_r405174295
 
 

 ##########
 File path: 
sql/core/src/test/scala/org/apache/spark/sql/streaming/StreamingAggregationSuite.scala
 ##########
 @@ -202,47 +202,68 @@ class StreamingAggregationSuite extends 
StateStoreMetricsTest with Assertions {
         }
       }
 
-      def stateOperatorProgresses: Seq[StateOperatorProgress] = {
-        val operatorProgress = mutable.ArrayBuffer[StateOperatorProgress]()
-        var progress = query.recentProgress.last
-
-        operatorProgress ++= progress.stateOperators.map { op => 
op.copy(op.numRowsUpdated) }
-        if (progress.numInputRows == 0) {
-          // empty batch, merge metrics from previous batch as well
-          progress = query.recentProgress.takeRight(2).head
-          operatorProgress.zipWithIndex.foreach { case (sop, index) =>
-            // "numRowsUpdated" should be merged, as it could be updated in 
both batches.
-            // (for now it is only updated from previous batch, but things can 
be changed.)
-            // other metrics represent current status of state so picking up 
the latest values.
-            val newOperatorProgress = sop.copy(
-              sop.numRowsUpdated + 
progress.stateOperators(index).numRowsUpdated)
-            operatorProgress(index) = newOperatorProgress
-          }
-        }
+      // Pick the latest progress that actually ran a batch
+      def lastExecutedBatch: StreamingQueryProgress = {
+        query.recentProgress.filter(_.durationMs.containsKey("addBatch")).last
+      }
 
-        operatorProgress
+      def stateOperatorProgresses: Seq[StateOperatorProgress] = {
+        lastExecutedBatch.stateOperators
       }
     }
 
+    val clock = new StreamManualClock()
+
     testStream(aggWithWatermark)(
+      // batchId 0
       AddData(inputData, 15),
-      CheckAnswer(), // watermark = 5
+      StartStream(Trigger.ProcessingTime("interval 1 second"), clock),
+      CheckAnswer(), // watermark = 0
       AssertOnQuery { _.stateNodes.size === 1 },
       AssertOnQuery { _.stateNodes.head.metrics("numOutputRows").value === 0 },
       AssertOnQuery { _.stateOperatorProgresses.head.numRowsUpdated === 1 },
       AssertOnQuery { _.stateOperatorProgresses.head.numRowsTotal === 1 },
+      AssertOnQuery { _.lastExecutedBatch.sink.numOutputRows == 0 },
+
+      // batchId 1 without data
+      AdvanceManualClock(1000L), // watermark = 5
+      Execute { q =>             // wait for the no data batch to complete
 
 Review comment:
   yeah, it'd be nice to provide an inbuilt function for it if this pattern is 
used more over time

----------------------------------------------------------------
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.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to