[
https://issues.apache.org/jira/browse/FLINK-7561?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16148744#comment-16148744
]
ASF GitHub Bot commented on FLINK-7561:
---------------------------------------
Github user pnowojski commented on the issue:
https://github.com/apache/flink/pull/4626
I have benchmarked this code and here are the results of using this
operator for tumbling windows, with 1000 number of distinct values in the key
and watermark once every 16000 elements:
Benchmark | Score | Error | Units
---------- | ----- | ----- | -----
WindowBenchmarks.preAggregateTumblingWindow | 4136.634 | ± 72.370 | ops/ms
WindowBenchmarks.tumblingWindow | 2592.292 | ± 67.317 | ops/ms
(benchmark code: https://github.com/dataArtisans/flink-benchmarks/pull/2 )
> Add support for pre-aggregation in DataStream API
> -------------------------------------------------
>
> Key: FLINK-7561
> URL: https://issues.apache.org/jira/browse/FLINK-7561
> Project: Flink
> Issue Type: New Feature
> Components: DataStream API
> Reporter: Piotr Nowojski
> Assignee: Piotr Nowojski
>
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
