CODEWITHSUNDEEP
pythonarrayspandasnumpyapply
Georg Heiler
Georg Heiler

Reputation: 17704

pandas extract array to columns

How can the array (the length of the array is constant for all elements in the series) be extracted into columns efficiently?

import pandas as pd
d = pd.DataFrame({'foo':\[1,2,3\], 'bar':\[\[1,1,1\], \[2,2,2\], \[3,3,3\]\]})
d][1]][1]

enter image description here

I.e. extract the array of [1,1,1] into a bar_0, bar_1, bar_3 column?

Is there a better way than manually iterating over the indices in the array and calling pandas.apply?

Upvotes: 2

Views: 1972

Answers (2)

Gusti Adli
Gusti Adli

Reputation: 1213

How about this:

>>> d.join(pd.DataFrame(d['bar'].to_list(), columns=['bar_1', 'bar_2', 'bar_3']))
   foo        bar  bar_1  bar_2  bar_3
0    1  [1, 1, 1]      1      1      1
1    2  [2, 2, 2]      2      2      2
2    3  [3, 3, 3]      3      3      3

You convert the bar column to list (nested list), convert it to a dataframe, and join the new dataframe with your initial dataframe.

Upvotes: 5

Andrej Kesely
Andrej Kesely

Reputation: 195553

import pandas as pd

d = pd.DataFrame({"foo": [1, 2, 3], "bar": [[1, 1, 1], [2, 2, 2], [3, 3, 3]]})

d = pd.concat([d, d.pop("bar").apply(pd.Series).add_prefix("bar_")], axis=1)
print(d)

Prints:

   foo  bar_0  bar_1  bar_2
0    1      1      1      1
1    2      2      2      2
2    3      3      3      3

Upvotes: 5

Related Questions