python - Dataframe manipulation, something with groupby, probably :/ -

need dataframe transformation. have

df = pd.dataframe({'c' : [1, 2,  1, 1, 2, 2,  1],                    'd' : [1, 2, 13, 4, 5, 9, 10]}) df = df.sort('c') =>       c   d      0   1   1      1   1  13      2   1   4      3   1  10      4   2   2      5   2   5      6   2   9 

and get

    f1   f2   f3   f4 1    1   13   4    10 2    2    5   9   nan 

currently, loop on everything, on large dataframe, it's slow.

thanks

you adding new f column , calling pivot:

>>> df["f"] = "f" + (df.groupby("c").cumcount() + 1).astype(str) >>> d2 = df.pivot(index="c", columns="f", values="d") >>> d2 f  f1  f2  f3  f4 c                 1   1  13   4  10 2   2   5   9 nan 

the f , c there index , column names; if want rid of set them none.

>>> d2.index.name = none >>> d2.columns.name = none >>> d2    f1  f2  f3  f4 1   1  13   4  10 2   2   5   9 nan