hi all In light of
https://github.com/apache/arrow/pull/9621 and some related issues, we're planning to start building out some query execution machinery in C++, with the goal of plumbing together the existing Datasets API with essential relational algorithms (projection, aggregation, filter, eventually join, sorting, etc.). We don't have SQL support planned for the time being so I would say this work is a good deal narrower in scope compared with the Rust DataFusion work, but we would like to have some high quality, reusable algorithms in the C++ core library and a lightweight multithreaded runtime to use them. Here's a document to help organize some of the initial labor and point out areas where we may need to build some new abstractions or refactor existing code related to Datasets: https://docs.google.com/document/d/1AyTdLU-RxA-Gsb9EsYnrQrmqPMOYMfPlWwxRi1Is1tQ/edit#heading=h.t89hffc3t7si Comments welcome. Thanks Wes