pandas.core.groupby.GroupBy.head

GroupBy.head(self, n=5)

Return first n rows of each group.

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

Returns:

Series or DataFrame

See also

Series.groupby

DataFrame.groupby

Examples

>>> df = pd.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