MiguelAnzoWizeline commented on pull request #14811: URL: https://github.com/apache/beam/pull/14811#issuecomment-912839733
Hi @lukecwik thanks for the comment. I have been thinking on the options you mentioned and I think that waiting for BEAM-12522 would be my preferred option over the CoGroupByKey one mainly to avoid changing the input to <Transaction, ReadOperation> as it seems to me a less natural way to approach it than obtaining the Transaction via SideInput and having to only process <ReadOperation>, but of course I could be wrong in this. Regarding your questions, I'm sorry but I currently don't really know how many partitions or output is generally produced when working with the Spanner IO, but if sub-partition splitting would help increase parallelism then I think the approach to take here would be to merge both DoFn into one DoFn<ReadOperation, Struct> where we take the transaction from the SideInput to create the partitions and split based on the partitions, this when BEAM-12522 is done. -- 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]
