How to load a FreqTableSheet-like file?

[frosencrantz]

It is possible to generate a summary sheet like a FreqTableSheet or a DescribeSheet outside of VisiData. For example using mlr to generate a file with FreqTableSheet like data:

mlr --csv stats1 -g Customer,Item -f Quantity -a sum benchmark.csv | vd

Screenshot: https://asciinema.org/a/krZwxKHFOp5QH40CqyIhZ2dTc

It is easy to have VisiData read this file, but some of the useful features of a FreqTableSheet or a DescribeSheet are the ability to dive into the data on the source sheet, or to update the selection on the source sheet.

I was wondering if I had the original file, and some sort of derived file that has some matching columns,how it is possible to easily get a similar behavior?

It would be useful to be able to load both files, and tell VisiData about the connection, in such a way that you could get the ability to dive into matching rows or update the selection on matching rows, just like it is possible with frequency and describe sheets. I know that as part of generating the FreqTableSheet, VisiData keeps lists of the source rows that match each Freq row, so the structures are not available just by loading a file.

There is a workaround, it is possible to create the FreqTableSheet and do a join, but this can be slow. It seems like it would be useful if it would be possible to dive without having to first bin all the rows. Though the dive would be much slower as it would need to find the matching rows.