Using column values as column names in dataframe

Question

I have the following dataframe:

df = pd.DataFrame([{'animal': 'dog', 'years':10}, 
                   {'animal': 'dog', 'years':5},
                   {'animal': 'cat', 'years':3},
                   {'animal': 'cat', 'years':7}])

giving me:

  animal  years
0    dog     10
1    dog      5
2    cat      3
3    cat      7

Given the number of cats and dogs is always the same, how can I can turn it into:

       dog   cat
years  10     3
        5     7

My ultimate goal is to make a boxplot showing years distrubution across each animal type.

Does this answer your question? [How to pivot a dataframe](https://stackoverflow.com/questions/47152691/how-to-pivot-a-dataframe) — villoro, Mar 06 '20 at 11:59
@villoro Unfortunately not. It lists pivot methods in pandas that seem to require aggregation of results. I do not need aggregation. jezrael's answer answers my question perfectly and leads straight to the point without going into lengthy explanations. — dzieciou, Mar 06 '20 at 12:46

score 2 · Accepted Answer · answered Mar 06 '20 at 11:23

Use GroupBy.cumcount for counter per animal and reshape by DataFrame.set_index and Series.unstack:

df1 = (df.set_index([df.groupby('animal').cumcount(), 'animal'])['years']
         .unstack()
         .rename_axis(None, axis=1))
print (df1)
   cat  dog
0    3   10
1    7    5

Using column values as column names in dataframe

1 Answers1