Hacker News Reader

Comment by vegabook | Hacker News Reader

I do a lot of...

  some_pandas_object.values

to get to the raw numpy, because often dealing with raw np buffer is more efficient or ergonomic. Hopefully losing numpy foundations will not affect (the efficiency of) code which does this.

View on HN

Looks like they're recommending using ".to_numpy()" instead of ".values" in dev (2.*) docs:

https://pandas.pydata.org/docs/dev/reference/api/pandas.Data...

.values has been deprecated for a long time now, with .to_numpy() as its recommended replacement for just as long.

Wow, had no idea about this. Hopefully they can just automatically interpret .values as .to_numpy or else a ton of old code is going to break.

That's what they've been doing for a long time as well!

I do the same. Sounds like arrow is being implemented as an alternative, and you'll have to explicitly opt in via a dtype or a global setting, so numpy isn't going away.

I often use Numba for complicated looping code for Pandas manipulation. Will be interesting to see how Arrow works with Numba.

… and what about interoperability with scikit-learn, PyTorch, etc?

None of that has changed.

Reply on news.ycombinator.com