> because there are a couple places where Avro, JSON are still hard-coded [2,3]
This is not a blocker, its due to the fact that PubsubTableProvider is just a wrapper for PubsubSchemaIOProvider. SchemaIOProvider requires you to specify all possible options, TableProvider does not. On Thu, Mar 18, 2021 at 2:44 PM Brian Hulette <bhule...@google.com> wrote: > Hi Fernando, > > Daniel Collins actually added the PayloadSerializerProvider concept very > recently [1], which is why it looks like Piotr's code doesn't apply > anymore. But the good news is I think that PR gets this task pretty close > to completion. It doesn't look like the PR *quite* finished adding support > for Proto to PubSubTableProvider though, because there are a couple places > where Avro, JSON are still hard-coded [2,3]. For this task to be complete > we should have tests of protobuf in PubSubTableProviderIT, which is > parameterized by payload type [3]. > > Regarding PubSubIO.readProtos: you're right to point out there's some > overlap between the various PubSub readers/writers and the TableProviders. > Ideally we'd define the logic for reading/writing Beam Rows with each IO in > a single place, but right now most of this logic lives in SQL's > TableProviders, and in a few places it's duplicated into the IOs, as with > readAvrosWithBeamSchema. For this task I think the right thing to do is use > PayloadSerializerProvider. > > [1] https://github.com/apache/beam/pull/13825 > [2] > https://github.com/apache/beam/blob/3fc2ab10d9f5d5c5b65ecf94ce45861857206674/sdks/java/io/google-cloud-platform/src/main/java/org/apache/beam/sdk/io/gcp/pubsub/PubsubSchemaIOProvider.java#L111 > [3] > https://github.com/apache/beam/blob/3fc2ab10d9f5d5c5b65ecf94ce45861857206674/sdks/java/extensions/sql/src/test/java/org/apache/beam/sdk/extensions/sql/meta/provider/pubsub/PubsubTableProviderIT.java#L108 > > On Thu, Mar 18, 2021 at 10:32 AM Fernando Morales Martinez < > fernando.mora...@wizeline.com> wrote: > >> Hi everyone, >> I'm working on the WI mentioned in the subject: "Add Proto support to >> Pubsub table provider" and I have a few questions . Sorry for the long mail! >> >> - The only method in KafkaTableProvider that performs some logic is >> buildBeamSqlTable. However, when taking a look at the tests in >> KafkaTableProviderIT, the only one that is calling that method appears to >> be testFake2. But if I’m not mistaken, that test doesn’t perform any test >> for the proto case. The only one that tests the proto case is testFake >> test, but that is only creating the KafkaTableProvider and that’s it. I >> wanted to base the PubSubTableProvider on the KafkaTableProvider since >> that >> one supports Proto and looks like it accomplishes that support by using >> PayloadSerializer. Is that correct? Should I follow that path? >> >> >> - I wanted to base the new code on the commits by Piotr, but a lot of >> the code he submitted appears to have been removed, but I can’t grasp the >> reason. There are several commits referencing one another and the Proto >> support. Can you shed some light on that? >> >> >> - Back to the PubsubTableProvider: PubSubIO, which contains >> reader/writer methods for protos (although via ProtoCoder), is being used >> by the PubsubSchemaIOProvider class for other read/write purposes. Why are >> the protos reader/writer not used? Because of the ProtoCoder? Should we >> instead implement new readers/writers by formatting the payload directly, >> in a similar fashion to readAvrosWithBeamSchema? >> >> Thanks a lot for the help! >> >> - Fernando Morales >> >> >> >> >> >> >> >> >> *This email and its contents (including any attachments) are being sent >> toyou on the condition of confidentiality and may be protected by >> legalprivilege. Access to this email by anyone other than the intended >> recipientis unauthorized. If you are not the intended recipient, please >> immediatelynotify the sender by replying to this message and delete the >> materialimmediately from your system. Any further use, dissemination, >> distributionor reproduction of this email is strictly prohibited. Further, >> norepresentation is made with respect to any content contained in this >> email.* > >
