GroupBy.head()

Returns first n rows of each group.

Essentially equivalent to .apply(lambda x: x.head(n)), except ignores as_index flag.

Examples

>>> df = DataFrame([[1, 2], [1, 4], [5, 6]],
                   columns=['A', 'B'])
>>> df.groupby('A', as_index=False).head(1)
   A  B
0  1  2
2  5  6
>>> df.groupby('A').head(1)
   A  B
0  1  2
2  5  6