dividing floating points into buckets

Question

I need to create a histogram of values that are floating points, and I'm wondering how exactly can I do that. This naive approach:

>>> 5.3/0.2
26.499999999999996
>>> 5.2/0.2
26.0

of dividing and then storing those in a dict will obviously not work.. round() is also not good enough since I want to have buckets of size 0.2. I could make buckets sized 0.1 and then combine them... Can anyone suggest an elegant way of doing this?

Answer 1

Use floored division to obtain the correct bin number:

>>> 5.3//0.2
26.0

Or on really old versions of python, you can do the same thing yourself using math.floor:

>>> math.floor(5.3 / 0.2)
26.0

In general, to calculate the bin number you could do something like this:

def get_bin(x, bin_width, start=0): 
    return (x - start) // bin_width

Where x is your number and start is the lower bound of the first bin.

As mentioned in the comments, you may also be interested in numpy.histogram.