[BEAM-10042] Add alternate constructor to improve byte encoding performance in SortValues

• Type: Improvement
• Status: Open
• Priority: P3
• Resolution: Unresolved
• Component/s: extensions-java-sorter
• Labels: Clarified
• Assignee: Unassigned
• Reporter: Claire McGinty
• Created: 20/May/20 13:29
• Updated: 04/Jun/22 16:26

Description

The `SortValues` transform operates on key-groups of `KV<PrimaryKeyT, Iterable<KV<SecondaryKeyT, ValueT>>>`. From those key groups it iterates through each element and uses `CoderUtils.encodeToByteArray` on each SecondaryKeyT-ValueT pair. This operation can be expensive and its parallelism is limited by the # of key groups.

I'd like to propose adding an alternative to `SortValuesDoFn` that operates on `KV<PrimaryKeyT, Iterable<KV<byte[], byte[]>>>` and can skip the encoding step within the key-group. The user's pipeline may be able to encode the data to bytes in a prior step in a much more parallelized and efficient way (i.e. in a `MapElements` transform). I've seen performance gains in every Dataflow metric from from patching this in my team's pipeline.

(I would visualize the alternative vs pre-existing constructors to look similar to generic vs specific Avro constructors, where the generic constructor has a static type and specific Avro has a parameterized T.)
 

What do you think? 

Issue Links

• links to: GitHub Pull Request #13059