From pandas

[TomAugspurger]

Taking Matt's idea

Are there benchmarks in Pandas that are appropriate to take?

Here's a bunch from some of https://github.com/pandas-dev/pandas/tree/master/asv_bench/benchmarks

All of these at least run. I need to go through and shorten the longer running ones.

[mrocklin]

Whoa, that's a lot of benchmarks :) I suspect that more-is-better generally, especially at this stage. Any thoughts on if/how we should organize into high or low priority?

[jorisvandenbossche]

@TomAugspurger I have a work-in-progress PR for cleaning up those benchmarks: pandas-dev/pandas#14099
I strongly recommend taking them from there or waiting until it is merged :-) (I am going to try to finish it this week). The original benchmarks were automatically generated based on the vbench ones and contain a huge amount of duplicate code and repetitive names.

I hope you didn't already do some of this clean-up yourself for this PR as that will probably have been duplicate work ..
(or if so, I can maybe pick some things from here for my PR)

[TomAugspurger]

@jorisvandenbossche 👍 not sure how I missed that. Most of the stuff I did was pep8 formatting, removing the * import, and excluding benchmarks that don't make sense for dask. I'll hold of till you merge that and re-submit this.