As we've learned, translating SQL to pandas isn't trivial. Additionally, there is a large group of potential Dask users who know SQL but would not consider themselves very proficient in pandas (myself included). To capture this audience and understand how good/bad their experience would be, it would be nice to add one or several projects that run SQL on Dask to the TPC-H benchmarks. dask-sql and ibis are top-of-mind here.
As we've learned, translating SQL to pandas isn't trivial. Additionally, there is a large group of potential Dask users who know SQL but would not consider themselves very proficient in
pandas
(myself included). To capture this audience and understand how good/bad their experience would be, it would be nice to add one or several projects that run SQL on Dask to the TPC-H benchmarks.dask-sql
andibis
are top-of-mind here.