Pandas groupby problem when dataframe has only one column

>>> import pandas as pd 
>>> import numpy as np
>>> df = pd.DataFrame(index=range(10),data=np.random.randint(0,3,10),columns=['y'])

Gives,

>>> df.y
0    2
1    0
2    2
3    1
4    0
5    0
6    0
7    1
8    2
9    1
Name: y, dtype: int32

>>> df.groupby('y').count()
Empty DataFrame
Columns: []
Index: [0, 1, 2]

But I'm trying to get the count of every item in 'y'. Something like,

>>> from collections import Counter
>>> Counter(df.y)
Counter({0: 4, 1: 3, 2: 3})

Am I doing something wrong here?

Thanks!

Comment From: jreback

The index is 'eaten' by the groupby so you need to pick a column to count.

In [6]: df.groupby('y')['y'].count()
Out[6]: 
y
0    1
1    5
2    4
Name: y, dtype: int64

you can also do this, treat as a series and groupby itself

In [11]: df['y'].groupby(df['y']).count()
Out[11]: 
y
1    4
2    6
Name: y, dtype: int64

Comment From: jorisvandenbossche

Or even better, you can just use value_counts():

In [22]: df['y'].value_counts(sort=False)
Out[22]:
0    3
1    5
2    2
Name: y, dtype: int64