boyuanzz commented on a change in pull request #11749:
URL: https://github.com/apache/beam/pull/11749#discussion_r445225947
##########
File path:
sdks/java/io/kafka/src/main/java/org/apache/beam/sdk/io/kafka/KafkaIO.java
##########
@@ -906,19 +1082,91 @@ public void setValueDeserializer(String
valueDeserializer) {
Coder<K> keyCoder = getKeyCoder(coderRegistry);
Coder<V> valueCoder = getValueCoder(coderRegistry);
- // Handles unbounded source to bounded conversion if maxNumRecords or
maxReadTime is set.
- Unbounded<KafkaRecord<K, V>> unbounded =
- org.apache.beam.sdk.io.Read.from(
-
toBuilder().setKeyCoder(keyCoder).setValueCoder(valueCoder).build().makeSource());
+ // The Read will be expanded into SDF transform when "beam_fn_api" is
enabled and
+ // "beam_fn_api_use_deprecated_read" is not enabled.
+ if (!ExperimentalOptions.hasExperiment(input.getPipeline().getOptions(),
"beam_fn_api")
+ || ExperimentalOptions.hasExperiment(
+ input.getPipeline().getOptions(),
"beam_fn_api_use_deprecated_read")) {
+ // Handles unbounded source to bounded conversion if maxNumRecords or
maxReadTime is set.
+ Unbounded<KafkaRecord<K, V>> unbounded =
+ org.apache.beam.sdk.io.Read.from(
+
toBuilder().setKeyCoder(keyCoder).setValueCoder(valueCoder).build().makeSource());
+
+ PTransform<PBegin, PCollection<KafkaRecord<K, V>>> transform =
unbounded;
+
+ if (getMaxNumRecords() < Long.MAX_VALUE || getMaxReadTime() != null) {
+ transform =
+
unbounded.withMaxReadTime(getMaxReadTime()).withMaxNumRecords(getMaxNumRecords());
+ }
- PTransform<PBegin, PCollection<KafkaRecord<K, V>>> transform = unbounded;
+ return input.getPipeline().apply(transform);
+ }
+ ReadAll<K, V> readTransform =
+ ReadAll.<K, V>read()
+ .withConsumerConfigOverrides(getConsumerConfig())
+ .withOffsetConsumerConfigOverrides(getOffsetConsumerConfig())
+ .withConsumerFactoryFn(getConsumerFactoryFn())
+ .withKeyDeserializerProvider(getKeyDeserializerProvider())
+ .withValueDeserializerProvider(getValueDeserializerProvider())
+ .withManualWatermarkEstimator()
+ .withTimestampPolicyFactory(getTimestampPolicyFactory());
+ if (isCommitOffsetsInFinalizeEnabled()) {
+ readTransform = readTransform.commitOffsets();
+ }
+ PCollection<KafkaSourceDescription> output =
+ input
+ .getPipeline()
+ .apply(Impulse.create())
+ .apply(ParDo.of(new GenerateKafkaSourceDescription(this)));
+ try {
+
output.setCoder(KafkaSourceDescription.getCoder(input.getPipeline().getSchemaRegistry()));
Review comment:
It works with `setSchema` but I want to make it explicitly because it's
possible that an user writes a DoFn which produces `KafkaSourceDescription`.
----------------------------------------------------------------
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:
us...@infra.apache.org
