4: Write a utility function to convert dollar values represented as strings (which do not sort correctly, e.g. '1B' sorts before '1k' and '10' sorts before '9') to dollar values represented as floats.

Write a function money_string_to_number(s) according to its docstring, below. You will need it, below, when processing DataFrame columns that contain money amounts like '3.4B' (3.4 billion dollars) and '7.8k' (7800 dollars).

In [ ]:

1

def money_string_to_number(s):

2

 """Returns float from money string s like '1.2T', '3.4B', '5.6M', or '7.8k'.

3

4

 To handle nan as an input, it first converts s to str(s).

5

 Then it makes these substitutions:

6

 'T' -> 'e12' (trillion)

7

 'B' -> 'e9' (billion)

8

 'M' -> 'e6' (million)

9

 'k' -> 'e3' (thousand)

10

 Then it uses float() to convert the string to float.

11

 """ 

 # ...

Here is some test code for the function:

inputs = ('1.2T', '3.4B', '5.6M', '7.8k')

outputs = [money_string_to_number(s) for s in inputs]

correct = [1.2e12, 3.4e9, 5.6e6, 7.8e3]

assert np.allclose(outputs, correct)

assert np.isnan(money_string_to_number(np.nan))