How to Remove Columns in Numpy Array that Contains Non-Numeric Values

When a NumPy numeric array contains NaN values, you often want to drop any column that has at least one NaN so the remaining array contains only fully numeric columns.

For Example: This example shows a simple case where a column containing NaN is removed from a 2×2 array.

import numpy as np
arr = np.array([[1.2, np.nan], [3.4, 5.6]])
print(arr[:, ~np.isnan(arr).any(axis=0)])

[[1.2]
 [3.4]]

Explanation: The removal of columns with non-numeric (NaN) values is done using np.isnan() + any(axis=0) + boolean indexing:

• np.isnan(arr): Creates a boolean array where NaN values are marked as True.
• .any(axis=0): Checks each column and marks it True if that column contains at least one NaN.
• ~ (NOT operator): Inverts the result so that only columns without NaNs are kept.
• arr[:, mask]: Selects and returns only the valid columns.

This way, all columns containing NaNs are removed in one clean step.

More Examples

Example 1: In this example, we remove columns containing non-numeric values from the 2X3 Numpy array.

import numpy as np

n_arr = np.array([[10.5, 22.5, np.nan], [41, 52.5, np.nan]])
print("Array:")
print(n_arr)

print("Result: ")
print(n_arr[:, ~np.isnan(n_arr).any(axis=0)])

Array:
[[10.5 22.5  nan]
 [41.  52.5  nan]]
Result: 
[[10.5 22.5]
 [41.  52.5]]

Example 2: Here, we remove columns containing non-numeric values from the 3X3 Numpy array.

import numpy as np

n_arr = np.array([[10.5, 22.5, 10.5], [41, 52.5, 25], [100, np.nan, 41]])
print("Array:")
print(n_arr)

print("Result: ")
print(n_arr[:, ~np.isnan(n_arr).any(axis=0)])

Array:
[[ 10.5  22.5  10.5]
 [ 41.   52.5  25. ]
 [100.    nan  41. ]]
Result: 
[[ 10.5  10.5]
 [ 41.   25. ]
 [100.   41. ]]

Example 3: In this example, we remove columns containing non-numeric values from the 5X3 Numpy array.

import numpy as np

n_arr = np.array( [[10.5, 22.5, 3.8],
                  [23.45, 50, 78.7],
                  [41, np.nan, np.nan],
                  [20, 50.20, np.nan],
                  [18.8, 50.60, 8.8]] )
print("Array:")
print(n_arr)

print("Result:")
print(n_arr[:, ~np.isnan(n_arr).any(axis=0)])

Array:
[[10.5  22.5   3.8 ]
 [23.45 50.   78.7 ]
 [41.     nan   nan]
 [20.   50.2    nan]
 [18.8  50.6   8.8 ]]
Result:
[[10.5 ]
 [23.45]
 [41.  ]
 [20.  ]
 [18.8 ]]