Hi Group, We want to compute rolling window percentile on top of group by results. Is there an option to expose serialized MergingDigest which is used by PercentileTDigest, similar to DistinctCountRawHLL ?

Seems like we don't (Only HLL and ThetaSketch seem to have it). Could you file an issue? Perhaps may be contribute even (should be easy)?

Sure, I can contribute. Fairly straight forward to do it.

Hi, thanks for reporting the issue. LinkedIn Pinot team is trying to implement the sliding window aggregation, and this is what we’re working on right now. We’ll send out the design doc soon

Thanks @User, could you point me to the issue to track this (or create if one doesn't exist)?

Here it is the generic one for sliding window aggregations: https://github.com/apache/pinot/issues/7213 We’ll put more details to the issue

Thanks @User Mind if we add the raw function in the meantime ?

@User please go ahead with adding the raw tdigest function in the mean while

I have the changes ready, could you please add me to the repo to create a private branch/raise a pr ? Creating a fork to raise the pr.

That was quick, thanks a lot. Added minor comments.

Thanks @User. Updated the pr. Could you check when you get a chance ?

Will do