Evalml: Drop support for python 3.5

Created on 2 Mar 2020  ·  8Comments  ·  Source: alteryx/evalml

Should we drop python 3.5 support?

Numpy is dropping (or has dropped?) 3.5 support.

Doing so would mean we can close #167, as long as we're satisfied that bug is specific to 3.5 and not masking something else.

@kmax12

task testing

Most helpful comment

I think Dask Distributed is the same.

All 8 comments

I think Dask Distributed is the same.

i'd say we're good to drop 3.5

I'm going to make a PR for this (#438)!

RE discussion in dev meeting, we don't know of any python 3.5 requirement

FWIW, it appears python 3.5 will get security fixes until Nov 2020. Numpy and pandas latest releases still claim to support 3.5. I'm still fine dropping now, just noting.

@dsherry one thing, the latest pandas release is 1.0.1 (you linked to 0.24.0) and doesn't support.

🤦‍♂️of course, sorry! Pandas 1.0.1 doesn't support python 3.5, as @jeremyliweishih originally commented.

Was this page helpful?
0 / 5 - 0 ratings