Karthik Velayutham

Looks like @prutskov left a PR open before he left. We should try to get the fix merged in if possible.

@mvashishtha @jbrockmendel I was thinking about this. I haven't confirmed this but I was wondering if `datetime` objects would give us some trouble here, esp when we get into timezones...

@jbrockmendel I don't have a minimum example I could show off the bat, but I was wondering if pandas.DatetimeTZDtype could cause some trouble here. I've had some problems before, but...

``` before after ratio [05bf659b] [9ebeed08] + 955±8μs 1.23±0.2ms 1.29 omnisci.benchmarks.TimeIndexing.time_iloc([10000, 10], 'numpy_array_take_all_values') + 14.9±2ms 18.5±3ms 1.24 benchmarks.TimeArithmetic.time_abs([250, 250], 1) - 200±60ms 79.0±2ms 0.39 benchmarks.TimeArithmetic.time_mode([250, 250], 1) ```

Closing this PR since it's no longer needed. Ray 2.0.0 has been officially released and we seem to be handling it well in CI.

Could you share the script you are using to profile? @jbrockmendel

@jbrockmendel Ah ok. Is it possible to at least get a minimum reproducible example?

@naren-ponder were you planning to add any tests to this PR?

@Garra1980 yes, @Billy2551 is going to be taking this on. It'll probably be in a new PR, but we will keep this one up for now.

In this vein, test-compat-win (engine dask, python 3.6) has been notoriously slow in CI for some time now.