To count the occurences of a value in a numpy array. This will work:
>>> import numpy as np >>> a=np.array([0,3,4,3,5,4,7]) >>> print np.sum(a==3) 2
The logic is that the boolean statement produces a array where all occurences of the requested values are 1 and all others are zero. So summing these gives the number of occurencies. This works for arrays of any shape or dtype.
There are two methods I use to count occurences of all unique values in numpy. Unique and bincount. Unique automatically flattens multidimensional arrays, while bincount only works with 1d arrays only containing positive integers.
>>> unique,counts=np.unique(a,return_counts=True) >>> print unique,counts # counts[i] is equal to occurrences of unique[i] in a [0 3 4 5 7] [1 2 2 1 1] >>> bin_count=np.bincount(a) >>> print bin_count # bin_count[i] is equal to occurrences of i in a [1 0 0 2 2 1 0 1]
If your data are numpy arrays it is generally much faster to use numpy methods then to convert your data to generic methods.