Skip to content

[BEAM-4783] Fix issues created in #6181. #7690

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
kyle-winkelman wants to merge 1 commit into from
Closed

[BEAM-4783] Fix issues created in #6181. #7690

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 12 additions & 8 deletions ...rs/spark/src/main/java/org/apache/beam/runners/spark/translation/TransformTranslator.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -133,14 +133,8 @@ public void evaluate(GroupByKey<K, V> transform, EvaluationContext context) {
WindowedValue.FullWindowedValueCoder.of(coder.getValueCoder(), windowFn.windowCoder());

// --- group by key only.
Long bundleSize =
context.getSerializableOptions().get().as(SparkPipelineOptions.class).getBundleSize();
Partitioner partitioner =
(bundleSize > 0)
? new HashPartitioner(context.getSparkContext().defaultParallelism())
: null;
JavaRDD<WindowedValue<KV<K, Iterable<WindowedValue<V>>>>> groupedByKey =
GroupCombineFunctions.groupByKeyOnly(inRDD, keyCoder, wvCoder, partitioner);
GroupCombineFunctions.groupByKeyOnly(inRDD, keyCoder, wvCoder, getPartitioner(context));

// --- now group also by window.
// for batch, GroupAlsoByWindow uses an in-memory StateInternals.
Expand Down Expand Up @@ -377,6 +371,7 @@ public void evaluate(
(KvCoder) context.getInput(transform).getCoder(),
windowingStrategy.getWindowFn().windowCoder(),
(JavaRDD) inRDD,
getPartitioner(context),
(MultiDoFnFunction) multiDoFnFunction);
} else {
all = inRDD.mapPartitionsToPair(multiDoFnFunction);
Expand Down Expand Up @@ -420,14 +415,15 @@ private static <K, V, OutputT> JavaPairRDD<TupleTag<?>, WindowedValue<?>> statef
KvCoder<K, V> kvCoder,
Coder<? extends BoundedWindow> windowCoder,
JavaRDD<WindowedValue<KV<K, V>>> kvInRDD,
Partitioner partitioner,
MultiDoFnFunction<KV<K, V>, OutputT> doFnFunction) {
Coder<K> keyCoder = kvCoder.getKeyCoder();

final WindowedValue.WindowedValueCoder<V> wvCoder =
WindowedValue.FullWindowedValueCoder.of(kvCoder.getValueCoder(), windowCoder);

JavaRDD<WindowedValue<KV<K, Iterable<WindowedValue<V>>>>> groupRDD =
GroupCombineFunctions.groupByKeyOnly(kvInRDD, keyCoder, wvCoder, null);
GroupCombineFunctions.groupByKeyOnly(kvInRDD, keyCoder, wvCoder, partitioner);

return groupRDD
.map(
Expand Down Expand Up @@ -550,6 +546,14 @@ public String toNativeString() {
};
}

private static Partitioner getPartitioner(EvaluationContext context) {
Long bundleSize =
context.getSerializableOptions().get().as(SparkPipelineOptions.class).getBundleSize();
return (bundleSize > 0)
? null
: new HashPartitioner(context.getSparkContext().defaultParallelism());
}

private static final Map<String, TransformEvaluator<?>> EVALUATORS = new HashMap<>();

static {
Expand Down
7 changes: 6 additions & 1 deletion ...ava/org/apache/beam/runners/spark/translation/streaming/StreamingTransformTranslator.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -83,6 +83,7 @@
import org.apache.beam.sdk.values.WindowingStrategy;
import org.apache.beam.vendor.guava.v20_0.com.google.common.collect.ImmutableMap;
import org.apache.spark.Accumulator;
import org.apache.spark.HashPartitioner;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.JavaSparkContext$;
Expand Down Expand Up @@ -305,7 +306,11 @@ public void evaluate(GroupByKey<K, V> transform, EvaluationContext context) {
JavaDStream<WindowedValue<KV<K, Iterable<WindowedValue<V>>>>> groupedByKeyStream =
dStream.transform(
rdd ->
GroupCombineFunctions.groupByKeyOnly(rdd, coder.getKeyCoder(), wvCoder, null));
GroupCombineFunctions.groupByKeyOnly(
rdd,
coder.getKeyCoder(),
wvCoder,
new HashPartitioner(rdd.rdd().sparkContext().defaultParallelism())));

// --- now group also by window.
JavaDStream<WindowedValue<KV<K, Iterable<V>>>> outStream =
Expand Down