Hi all,

Although the Druid 38 release is still in progress, I’d like to start
discussing two feature areas I’m hoping to complete for Druid 39.
Both have foundational PRs that would benefit from review now, since the
follow-up work builds on them.

- #20183 <https://github.com/apache/druid/pull/20183> — Add native query
support for system tables establishes an opt-in native query path for
supported system tables, with sys.server_properties as the initial
implementation.
With this support, performance of point queries on sys.tasks and analytic
queries on sys.segments can be maximized, and ALL native functions can be
applied to system tables.
I expect roughly 6 follow-up PRs to extend support to other tables,
including sys.tasks and sys.segments.
The bottom line here is that sys.tasks and sys.segments are included in the
native query path.

- #20198 <https://github.com/apache/druid/pull/20198>— Add typed query
context parameters introduces a centralized, strongly typed model for query
context parameters.
It migrates two parameters as a starting point and generates their
documentation from code. Only with this support can we validate a parameter
in a SET statement.
I expect another 3 or 4 PRs to migrate the remaining parameters and add a
sys.query_context_parameters system table.
The bottom line here is to complete all parameter migration and have the
SET statement validated.

I’d appreciate reviews and feedback on these foundational PRs, especially
any design concerns or blockers that could affect the follow-up work.
Please also share your thoughts on the scope and timing for Druid 39.

Thanks and regards.

Reply via email to