In this article, we are going to see how to count the number of non-NaN elements in a NumPy array in Python.
NAN: It is used when you don’t care what the value is at that position. Maybe sometimes is used in place of missing data, or corrupted data.
Method 1: Using Condition
In this example, we will use one-dimensional arrays. In the below-given code, we loop over every entry of the given NumPy array and check if the value is a NaN or not.
Python3
import numpy as npex1 = np.array([1, 4, -9, np.nan])ex2 = np.array([1, 45, -2, np.nan, 3, -np.nan, 3, np.nan])def approach_1(data): # here the input data, is a numpy ndarray # initialize the number of non-NaN elements # in data count = 0 # loop over each entry of the data for entry in data: # check whether the entry is a non-NaN value # or not if not np.isnan(entry): # if not NaN, increment "count" by 1 count += 1 return countprint(approach_1(ex1))print(approach_1(ex2)) |
Output:
3 5
Method 2: Using isnan()
Using the functionality of NumPy arrays, that we can perform an operation on the whole array at once, instead of a single element.
Used function:
- np.isnan(data): Returns a boolean array after performing np.isnan() operation on one of the entries of the array, data
- np.sum(): Since we are inputting a boolean array to the sum function, it returns the number of True values (1s) in the bool array.
Python3
import numpy as npex3 = np.array([[3, 4, -390, np.nan], [np.nan, np.nan, np.nan, -90]])def approach_2(data): return np.sum(~np.isnan(data))print(approach_2(ex3)) |
Output:
4
Method 3: Using np.count_nonzero() function
numpy.count_nonzero() function counts the number of non-zero values in the array arr.
Syntax : numpy.count_nonzero(arr, axis=None)
Parameters :
arr : [array_like] The array for which to count non-zeros.
axis : [int or tuple, optional] Axis or tuple of axes along which to count non-zeros. Default is None, meaning that non-zeros will be counted along a flattened version of arr.Return : [int or array of int] Number of non-zero values in the array along a given axis. Otherwise, the total number of non-zero values in the array is returned.
Python3
import numpy as npex4 = np.array([[0.35834379, 0.67202438, np.nan, np.nan, np.nan, 0.47870971], [np.nan, np.nan, np.nan, 0.08113384, 0.70511741, 0.15260996], [0.09028477, np.nan, 0.16639899, 0.47740582, 0.7259116, 0.94797347], [0.80305651, np.nan, 0.67949724, 0.84112054, 0.15951702, 0.07510587], [0.28643337, 0.00804256, 0.36775056, 0.19360266, 0.07288145, 0.37076932]])def approach_3(data): return data.size - np.count_nonzero(np.isnan(data))print(approach_3(ex4)) |
Output:
22
