mosche opened a new issue, #24035: URL: https://github.com/apache/beam/issues/24035
### What would you like to happen? In the Spark dataset runner, side inputs are currently materialized and broadcasted per consuming `ParDo`. If a `PCollection` serves as side-input to multiple `ParDo`s, it would be evaluated, collected on the driver and finally broadcasted repeatedly for every such `ParDo`. ### Issue Priority Priority: 2 ### Issue Component Component: runner-spark -- 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]
