The current naming / import of the package feels inconsistent and not what I expect as a Python developer - especially if we're trying to reduce the barrier of entry from Pandas to Spark. It's not a big deal, but changing this could reduce friction.
I'd recommend being consistent (and shadowing Pandas) for both the pip install and import statements:
_Pandas style:_
pip install pandas
import pandas as pd
_Current Koalas style:_
pip install koalas
import databricks.koalas as ks
_Recommended Koalas style:_
pip install koalas
import koalas as ks
Thanks. I'm going to leave this ticket open here so others can chime in. We can consider it if there's enough request.
actually I personally like this way too but I think it's not a big deal.
Agree with all of the above.
I am +1 to the package koalas instead of databricks.koalas, mostly because other packages (pyspark, numpy, pandas, etc) looks that way
+1 on this
Also, using koalas as kl would be more consistent with pandas as pd and numpy as np.
Actually, let's don't change it for now unless there's technical reasons. It's troublesome to change it at the current status because we should move the packages around and users have to change their import style ...
Most helpful comment
I am +1 to the package
koalasinstead ofdatabricks.koalas, mostly because other packages (pyspark,numpy,pandas, etc) looks that way