Find differences between arrays, with probability of variation

Question

Good afternoon guys, I’m reading about the numpy library setdiff1d function (https://docs.scipy.org/doc/numpy/reference/generated/numpy.setdiff1d.html).

It basically serves to express the different values between 2 arrays, for example (found in the link itself):

>>> a = np.array([1, 2, 3, 2, 4, 1])
>>> b = np.array([3, 4, 5, 6])
>>> np.setdiff1d(a, b)
array([1, 2])

I would like to know if you have any function to insert a percentage of variation in the values to be compared.

I explain, as shown in the example, the setdiff1d function will return the values that are exactly different between the 2 arrays.

However, if I’m working with floats, and I want to allow a percentage of variation between these values, for example, considering 3.35 and 3.34 between two arrays as equal values, has some specific function to work in this way, where I can set as argument a margin of variation between the values to be compared?

grateful.

Answer 1

Try it this way:

limit = 0.01
np.unique(a[(np.abs(np.subtract.outer(a,b)) > limit).all(1)])