Fidel Fidel - 8 months ago 100
Python Question

How to count values in a certain range in a Numpy array? (Does not work)

I have already checked this thread How to count values in a certain range in a Numpy array? but their answer does not seem to work.

I have a numpy array of 2000 floats called data:

print(type(data)) --> <type 'list'>
print(type(data[0])) --> <type 'numpy.float64'>

And I have 2 variables to form a range, minV and maxV:

print(type(minV)) --> <type 'float'>
print(type(maxV)) --> <type 'float'>

If I try the solution given in the link mentioned above, I receive this exception:

((minV < data) & (data < maxV)).sum()

AttributeError: 'bool' object has no attribute 'sum'

And indeed, that expression is a boolean:

print(type( (minV < data) & (data < minV) ) ) --> <type 'bool'>
print( ( (minV < data) & (data < minV) ) ) --> True

The python version I am using is Python 2.7.3 -- EPD 7.3-2 (64-bit)
Numpy version is 1.6.1

System is Linux (Although I ignore if that is important).



I suspect you are using python 2 because comparing a list and a number doesn't raise a TypeError in your case.

But in order to use element-wise comparison (<, >, &) you need to convert your list to a numpy array:

import numpy as np
data = np.array(data)
((minV < data) & (data < maxV)).sum()

should work. For example:

data = list(range(1000))
minV = 100
maxV = 500
data = np.array(data)
((minV < data) & (data < maxV)).sum() # returns 399