search

apache / pinot

Apache Pinot - A realtime distributed OLAP datastore
https://pinot.apache.org/
Apache License 2.0
5.42k stars 1.27k forks source link

FILTER Clause Support #7519

Closed atris closed 1 year ago

atris commented 2 years ago

FILTER clause (SELECT COUNT(*) FILTER (WHERE foo > bar) FROM fooBar) allows complex reports to be generated efficiently without recomputation of the core predicate and rescans of the original data set.

https://docs.google.com/document/d/1qJzG1CmyVZpLN6rdI8V0zUWEaIjQ86hDL9XwrcPKcTs/edit?usp=sharing

siddharthteotia commented 2 years ago

Thanks @atris . Will review the doc

xiangfu0 commented 2 years ago

@yupeng9

Jackie-Jiang commented 2 years ago

Reference to the query syntax: https://modern-sql.com/feature/filter Postgres supports it: https://medium.com/little-programming-joys/the-filter-clause-in-postgres-9-4-3dd327d3c852

Jackie-Jiang commented 2 years ago

The aggregation only support is merged #7916

We should support group-by as the next step

egalpin commented 2 years ago

Is there a known path/existing plan for how to implement group-by support? Seems interesting to work on if it were Pinot-dev-beginner friendly

atris commented 2 years ago

I am currently working on a PR for the same (Side note - it is not a beginner friendly PR)

On Sat, 21 May 2022, 05:10 Evan Galpin, @.***> wrote:

Is there a known path/existing plan for how to implement group-by support? Seems interesting to work on if it were Pinot-dev-beginner friendly

— Reply to this email directly, view it on GitHub https://github.com/apache/pinot/issues/7519#issuecomment-1133466496, or unsubscribe https://github.com/notifications/unsubscribe-auth/AANE5Y47N6DXIZXBV4VMR73VLAPH3ANCNFSM5FLB2EPA . You are receiving this because you were mentioned.Message ID: @.***>

egalpin commented 1 year ago

@atris Any update on the progress of a PR to support GROUP BY? Would be a very exciting feature to see 😊

egalpin commented 1 year ago

I wanted to mention that I'm currently exploring/investigating GROUP BY support with FILTER expressions. As best as I can tell so far, GroupByPlanNode will need to have similar modifications as those made to AggregationPlanNode in @atris original PR (plus all implicated changes). I realize it's been advised that this is not a beginner-friendly PR but support for FILTER + GROUP BY will be high impact for my use of Pinot, so it's worth a shot.

I welcome any collaboration!