Zhang Mingli
Zhang Mingli
> we generated two different plans for those two SQL statements which are functionally equivalent. I debug this for a while: For CTE SQL, the HashAgg on SeqScan came from...
> For SQL: `select count(distinct id) from ao;` > I found we already have the codes to add an additional DEDUP HashAgg path in PG planner. > Yes, there was...
Hi all, fixed in #676 ```sql explain(costs off) select count(distinct a) from t_issue_659; QUERY PLAN ------------------------------------------------- Finalize Aggregate -> Gather Motion 3:1 (slice1; segments: 3) -> Partial Aggregate -> HashAggregate...
> Does it only work when the key of distinct is the same as the distribution key, or it doesn't matter? GPDB already have 3-phase agg plan when distinct key...
This is a cherry-pick commit from commercial version.
> What I cannot understand right now - how to use bloom filter in MPP environment? Is it enough to create local bloom filters? HI, do you mean how to...
> > Do we have a **Parallel Plan** test(ex: cbdb_parallel.sql) to ensure the output is as expected? > > I'm not sure what we need **Parallel Plan** test, but I...
@fanfuxiaoran Thanks for your analysis and verification. > > Do we need to add the parallel worker info into this line? like > > ``` > Rows out: 333333.33 rows...
Another flaky failed cases of resgroup: https://github.com/cloudberrydb/cloudberrydb/issues/651#issuecomment-2387565620
Nestloop tends to use normal plan instead of Parallel, need a dig.