Convert a pandas groupby object to list of dataframes

Question

Say I have the following dataframe, and want to group-by the ys:

I can do this by running

grouped = df.groupby('ys')

I can iterate through this new groupby object fine, but instead I want a list of the dataframes that are accessed by group in the following loop:

for name, group in grouped:
    do_something(group)

Is this possible?

You want a list that is accessed by a *group name*? That's impossible. You probably want a `dict`. — juanpa.arrivillaga, Jan 16 '17 at 17:26
No, I just want a list of the dataframes and don't care about the names. Just realised my loop was wrong, should be corrected now — David Stansby, Jan 16 '17 at 17:28

score 26 · Answer 1 · answered Jan 16 '17 at 17:29

26

Sure, just iterate over the groups!

>>> import pandas as pd, numpy as np
>>> df = pd.DataFrame(dict(xs=list(range(4)), ys=[0,0,1,1]))
>>> df
   xs  ys
0   0   0
1   1   0
2   2   1
3   3   1
>>> grouped = df.groupby('ys')
>>> dataframes = [group for _, group in grouped]
>>> dataframes
[   xs  ys
0   0   0
1   1   0,    xs  ys
2   2   1
3   3   1]
>>>

answered Jan 16 '17 at 17:29

juanpa.arrivillaga

88,713
10
131
172

Of course... almost answered myself in the question! Thanks for the answer! – David Stansby Jan 16 '17 at 17:32
can smn explain [group for _, group in grouped] how it works? – bohontw Nov 10 '21 at 18:31
1

@bcikili, it is equivalent to [group for name, group in grouped]. The underscore is used to emphasize that we don't care about the name. – Evgenii Nov 19 '21 at 15:37

Convert a pandas groupby object to list of dataframes

1 Answers1

Linked

Related