Use the Series rank
method:
In [11]: df.a.rank()
Out[11]:
0 4
1 1
2 8
3 10
4 6
5 2
6 3
7 9
8 7
9 5
Name: a, dtype: float64
It has a correspinding ascending argument:
In [12]: df.a.rank(ascending=False)
Out[12]:
0 7
1 10
2 3
3 1
4 5
5 9
6 8
7 2
8 4
9 6
Name: a, dtype: float64
In the case of ties, this will take the average rank, you can also choose min, max or first:
In [21]: df = pd.DataFrame(np.random.randint(1, 5, (10, 2)), columns=list('ab'))
In [22]: df
Out[22]:
a b
0 2 2
1 3 4
2 1 1
3 3 1
4 4 2
5 2 4
6 1 4
7 2 1
8 1 2
9 3 4
In [23]: df.a.rank() # there are several 2s (which have rank 5)
Out[23]:
0 5
1 8
2 2
3 8
4 10
5 5
6 2
7 5
8 2
9 8
Name: a, dtype: float64
In [24]: df.a.rank(method='first')
Out[24]:
0 4
1 7
2 1
3 8
4 10
5 5
6 2
7 6
8 3
9 9
Name: a, dtype: float64