What Series method replaced searchsorted?

Question

In his video, [Data analysis in Python with pandas] (http://youtu.be/w26x-z-BdWQ?t=2h14s), Wes McKinney presents a series method names searchsorted(), which given a value, gives back the index in which the series is crossing that value. It appears this function is not available any more, did something else replace it?

Answer 1

I believe this is due to the refactoring that occurred in Pandas 0.13.0 where Pandas Series now sub-class NDFrame rather than ndarray see this:

In [33]:

import pandas as pd
import numpy as np
df = pd.DataFrame({'a':arange(10)})
df
Out[33]:

   a
0  0
1  1
2  2
3  3
4  4
5  5
6  6
7  7
8  8
9  9

[10 rows x 1 columns]

[10 rows x 3 columns]
In [28]:

# you now have to call `.values` to return a ndarray 
df.a.values.cumsum().searchsorted(11)
Out[28]:
5

Now compare what happens if we use a numpy array:

In [29]:

temp = np.array(arange(10))

In [32]:

temp.cumsum().searchsorted(11)
Out[32]:
5