Get the first row of each group of unique values in another column

Question

I want to extract from a dataframe only the first row of each value other than one of the columns (pandas) for example:

df
   col_A col_B
0      1     x
1      2    xx
2      3    xx
3      4     y
4      5     y

to

df1
  col_A col_B
0      1     x
1      2    xx
2      4     y

Answer 1

Use groupby + first:

firsts = df.groupby('col_B', as_index=False).first()

Output:

>>> firsts
  col_B  col_A
0     x      1
1    xx      2
2     y      4

If the order of the columns is important:

firsts = df.loc[df.groupby('col_B', as_index=False).first().index]

Output:

>>> firsts
   col_A col_B
0      1     x
1      2    xx
2      3    xx