[
https://issues.apache.org/jira/browse/HUDI-8301?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Geser Dugarov updated HUDI-8301:
--------------------------------
Description: Currently `StreamWriteFunction` uses buffer with
`DataBucket`s, which contains `DataItem`s transformed `HoodieRecord`s. We could
use `HoodieRecord`s directly to reduce number of objects in the heap, and
decrease GC workload.
> Remove DataItem using during Flink stream write
> -----------------------------------------------
>
> Key: HUDI-8301
> URL: https://issues.apache.org/jira/browse/HUDI-8301
> Project: Apache Hudi
> Issue Type: Improvement
> Reporter: Geser Dugarov
> Assignee: Geser Dugarov
> Priority: Major
>
> Currently `StreamWriteFunction` uses buffer with `DataBucket`s, which
> contains `DataItem`s transformed `HoodieRecord`s. We could use
> `HoodieRecord`s directly to reduce number of objects in the heap, and
> decrease GC workload.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
