Hi Beam Community,

I'd like to add read support to the existing Java ClickHouseIO connector
and get feedback on the proposed design before starting implementation.

The initial version would support bounded SQL reads, parallel table reads,
and Beam Row output. readAll() and advanced partitioning features are
deferred to future work.

Design doc:
https://docs.google.com/document/d/1oxBZg8oZVfsgYcMkRagdNekiS58GMkbH/edit

Tracking issue: https://github.com/apache/beam/issues/40048

I've included a few open questions in the design doc, particularly around
parallel reads, partitioning, and schema handling.

I'd appreciate feedback on these or any other aspects of the design.

Thanks,
Nitin Ware (@nitinware <https://github.com/nitinware>)

Reply via email to