KevinGG commented on issue #21430: URL: https://github.com/apache/beam/issues/21430#issuecomment-1211022613
It just skips the pruning for a specific transform, not the whole pruning feature. TL;DR: the pruning logic is pretty naive, it marks all transforms and input/ouptut PCollections contributing to a output PCollection bottom-up until it hits the root transform or cached PCollections. The problem with the pruning logic is that it can preserve only partial transforms/PCollections from a composite transform, which can be a problem (maybe some context is shared within a composite but not in the format of a transform/pcollection). The mitigation is to skip the pruning of such composite transforms. The solution might require rewriting the composite transform or add a special composite pruning logic for it. -- 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]
