[BEAM-7448] Remove redundant windowing information and unused accumulators on Spark runner

• Type: Improvement
• Status: Resolved
• Priority: P2
• Resolution: Fixed
• Fix Version/s: 2.14.0
• Component/s: runner-spark
• Assignee: Etienne Chauchot
• Reporter: Etienne Chauchot
• Created: 29/May/19 09:03
• Updated: 16/May/20 13:38
• Resolved: 29/May/19 09:37

Description

In some places the Spark runner uses both external windowing and internal windowing in collections such as

WindowedValue<KV<K, Iterable<WindowedValue<V>>>>

leading to unneeded mappers and dereference
The external windowing information is duplicated so unneeded

Issue Links

• links to: GitHub Pull Request #8711