Amar3tto commented on code in PR #34253:
URL: https://github.com/apache/beam/pull/34253#discussion_r1991738026
##########
runners/spark/src/main/java/org/apache/beam/runners/spark/translation/SparkStreamingPortablePipelineTranslator.java:
##########
@@ -322,13 +342,21 @@ private static <T> void translateFlatten(
dStreams.add(unboundedDataset.getDStream());
} else {
// create a single RDD stream.
- final SingleEmitInputDStream<WindowedValue<T>>
singleEmitInputDStream =
- new SingleEmitInputDStream<WindowedValue<T>>(
- context.getStreamingContext().ssc(), ((BoundedDataset)
dataset).getRDD().rdd());
- final JavaDStream<WindowedValue<T>> dStream =
- JavaDStream.fromDStream(
- singleEmitInputDStream,
JavaSparkContext$.MODULE$.fakeClassTag());
-
+ JavaDStream<WindowedValue<T>> dStream;
+ if (context.getSparkContext().version().startsWith("3")) {
+ Queue<JavaRDD<WindowedValue<T>>> q = new LinkedBlockingQueue<>();
+ q.offer(((BoundedDataset) dataset).getRDD());
+ // TODO (https://github.com/apache/beam/issues/20426): this is not
recoverable from
Review Comment:
#34080 should still work for Spark 2 / Non-portable runner. I guess that PR
was made and tested only for Spark 2.
Spark 3 test was failing, so my PR should fix it.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
