Johan Adami

For those looking for an easy way to test this in a notebook, you can clear the ruleset directory before recreating the rule. I use this to keep altering the...

note that this table has time and partition pruning. It seems when all segments are pruned, you get back no rows. But if not all segments are pruned, then you...

I tested locally with postgres's internal tables. It always returns a single row with null ``` postgres=# select max(stawidth) from pg_catalog.pg_statistic; max ------ 1237 (1 row) postgres=# select max(stawidth) from...

apologies missed this. Maybe the multi-stage engine supports this, but these are not valid queries at the moment in Pinot, so I don't think they apply. ``` SELECT * FROM...

I'm honestly not sure what the semantics would be here. I'm thinking more about the semantics of single aggregations. They shouldn't just return the default value if no rows match;...

Oh yes. Totally agree with those. Thank you!

cc @Jackie-Jiang since you've touched a lot of the pruner code recently. Does it make sense to have the pruner do ``` try { prune() } catch { return null...

between the new/old formats, https://docs.pinot.apache.org/configuration-reference/schema#new-datetime-formats, `1:MILLISECONDS:TIMESTAMP` isn't mentioned anywhere. Though if I read it correctly, just `TIMESTAMP` should work as you said above? Interestingly, this was/is being used as the...

also to provide more concrete feedback for the schema docs, https://docs.pinot.apache.org/configuration-reference/schema, there are ~1700 words on that page and all of 1 schema with 6 columns worth of examples. The...

> In `waitForJobIdToPersist()`, we can consider passing in the future and stop waiting if the future is done. This way we can avoid unnecessary wait because job id will never...