How do I pass in pre-grouped (+ aggregated) data?

[githorse]

It seems I can dramatically improve the performance (by 10x) of grouping and aggregation by doing it myself, at least for some common use cases¹. (Try this benchmark demo with 100k+ records to see the difference.) It's worth it to me to incur some extra complexity to do my own grouping and aggregation -- but I'm not sure how I would use this with Tanstack. @KevinVandy of Material React Table says this is possible, but I'm not sure how it would work. Can anyone suggest what this might look like in Tanstack?

(This technique is possible with AG Grid and presumably some other datagrids as well.)

Footnotes

1. It's also surprisingly simple with a solution like Alasql (again, for straightforward use cases). ↩