So far, I have seen one difference. The pandas set_option function throws on passing in a -1 as is done in the custom dataframe formatter in rf_ipython submodule. It should either be non-negative or None. Versions 0.2x. require the value passed to be type int.
I am not sure what our strategy should be, either to support both versions as soon as we can, or make a clean break to 1.0 at a known version, perhaps RF 1.0?
It's not clear to me how much traction pandas 1.0 is getting.
So far, I have seen one difference. The pandas
set_option
function throws on passing in a -1 as is done in the custom dataframe formatter inrf_ipython
submodule. It should either be non-negative orNone
. Versions 0.2x. require the value passed to be typeint
.I am not sure what our strategy should be, either to support both versions as soon as we can, or make a clean break to 1.0 at a known version, perhaps RF 1.0?
It's not clear to me how much traction pandas 1.0 is getting.