pandas.core.groupby.GroupBy.head¶
-
GroupBy.
head
(n=5)[source]¶ 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