Pandas: inverse of value_counts function

Question

What's a quick way to produce the inverse output of the value_counts function?

For example, if I have the following series:

1      24
2       2
3       1
4       2
5       3
6      12
7      21
8     204
9     400
10     71
11    160
Name: foo, dtype: float64

How can I concisely produce the following array?

numpy.array([1, 1, 1, ... , 2, 2, 3, 4, 4, 5, 5, 5, 6, ... ])

Answer 1

You can use np.repeat. If your Series is named s, it's possible to write:

np.repeat(s.index.values, s.values)

Here s.index.values are the values to repeat, and s.values specifies the number of times that each value should be repeated. The output is a 1D array.