1

I have a pandas dataframe with the following form:

import pandas as pd
p = pd.DataFrame({"int" : [1,     1,     1,     1,     2,      2],
                  "cod" : [[1,1], [2,2], [1,2], [3,9], [2,2], [2,2]]})

I want to group by int, which gives me a bunch of lists. I then want to flatten these lists, so I ultimately end up with a dataframe that has this form:

p = pd.DataFrame({"int" :  [1,                2],
                  "cod" : [[1,1,2,2,1,2,3,9], [2,2,2,2]]})

Here is what I have so far:

p.groupby("int", as_index=False)["cod"]

I'm stuck at how to flatten once I have grouped by int

N08
  • 1,265
  • 13
  • 23

1 Answers1

6

Use sum:

df = p.groupby("int", as_index=False)["cod"].sum()

Or list comprehension:

df = p.groupby("int")["cod"].apply(lambda x: [z for y in x for z in y]).reset_index()

df = p.groupby("int")["cod"].apply(lambda x: np.concatenate(x.values).tolist()).reset_index()

For performance if large list should be fastest:

from itertools import chain

df = p.groupby("int")["cod"].apply(lambda x: list(chain.from_iterable(x))).reset_index()

Check more information about flattening lists.

print (df)
   int                       cod
0    1  [1, 1, 2, 2, 1, 2, 3, 9]
1    2              [2, 2, 2, 2]
jezrael
  • 822,522
  • 95
  • 1,334
  • 1,252