Matthew Zeitlin

Copying from #238: after updating to `0.15.0`, a set of features that built fine in `0.14.1` now doesn't compile. A minimal `Cargo.toml` which compiles for `0.14` but not `0.15.0`: ```...

Thanks for reporting this @cdeil! Not sure if there are issues the `pandas` logic somewhere as well, but I can reproduce the segfault with only `numpy` with ``` td_arr =...

Issue fixed upstream (very quickly!) - we should probably add a test, but need to wait until 1.21.3 released.

There's some handling for name clashes here: https://github.com/pandas-dev/pandas/blob/42adb9f80fb6a048995087276eda2a16ea1cfb66/pandas/core/computation/engines.py#L27-L43 Contributions welcome to figure out if this occurs for other column names and where `pandas` might detect this and raise a more...

Related: #35595

> Yes @mzeitlin11 There is a issue #46646. Before this it was showing lot of error statements after this it just shows a few of them I believe Thanks. Can...

Thanks for reporting this @cynddl! While certainly confusing, I believe this is intended behavior. For `object` type, `pd.NA` and `np.nan` are both valid missing values and treated as distinct. For...

@TomAugspurger I would be interested in contributing, is it all right if I work on DataFrame.unstack?

Thanks for writing these examples up so clearly here @simonjayhawkins, made it very easy to get started! A potentially useful addition is prepending `CFLAGS="-O0"` to the run command, which made...

> Is this only true for bisecting? Any reason to not use it all the time? > Gives a tradeoff between compile time and performance. No optimizations gives much slower...