How can I check whether a numpy array is empty or not?

numpy

python

array

check-array

programming-tips

How can I check whether a numpy array is empty or not?

Master System Design with Codemia

Enhance your system design skills with over 120 practice problems, detailed solutions, and hands-on exercises.

Start Practicing Learn More

Introduction

A NumPy array is "empty" when it has zero elements (size == 0). This can happen with explicit empty arrays (np.array([])), arrays with zero-length dimensions (np.empty((0, 3))), or as the result of filtering operations that match nothing. The recommended way to check is array.size == 0. Avoid using Python's len() or truthiness checks (if array:) because they behave differently for multi-dimensional arrays and NumPy intentionally raises an error for ambiguous boolean evaluation.

Checking with `.size`

python

1import numpy as np
2
3# Empty array
4arr = np.array([])
5print(arr.size)    # 0
6print(arr.size == 0)  # True
7
8# Non-empty array
9arr = np.array([1, 2, 3])
10print(arr.size)    # 3
11print(arr.size == 0)  # False
12
13# Multi-dimensional empty array
14arr = np.empty((0, 5))
15print(arr.shape)   # (0, 5)
16print(arr.size)    # 0 — zero elements even though shape has 2 dimensions
17print(arr.size == 0)  # True
18
19# Array with a single zero is NOT empty
20arr = np.array([0])
21print(arr.size)    # 1
22print(arr.size == 0)  # False — it has one element (the value 0)

Why Not Use `len()`?

python

1import numpy as np
2
3# len() returns the size of the first dimension, not total elements
4arr = np.empty((0, 3))
5print(len(arr))      # 0 — happens to work here
6
7arr = np.empty((3, 0))
8print(len(arr))      # 3 — WRONG! Array has 0 elements but len() says 3
9
10arr = np.array([1, 2, 3])
11print(len(arr))      # 3 — correct for 1D
12
13# 0-dimensional array
14arr = np.array(42)
15# len(arr) raises TypeError: len() of unsized object

len() measures the first axis only. For multi-dimensional arrays or scalar arrays, it gives misleading or broken results. Always use .size.

Why Not Use `if array:`?

python

1import numpy as np
2
3arr = np.array([1, 2, 3])
4
5# This raises an error in NumPy
6if arr:
7    print("not empty")
8# ValueError: The truth value of an array with more than one element is ambiguous.
9# Use a.any() or a.all()
10
11# Single-element array works but is misleading
12arr = np.array([0])
13if arr:
14    print("not empty")
15# Does NOT print — the single element is 0 (falsy), but the array is NOT empty

NumPy intentionally makes boolean evaluation ambiguous for multi-element arrays. Do not use if array: to check emptiness.

Checking After Filtering

python

1import numpy as np
2
3data = np.array([10, 20, 30, 40, 50])
4
5# Filter: values greater than 100
6result = data[data > 100]
7print(result)         # []
8print(result.size)    # 0
9
10if result.size == 0:
11    print("No values matched the filter")
12else:
13    print(f"Found {result.size} values: {result}")

Checking for None vs Empty

python

1import numpy as np
2
3def process_data(arr):
4    # Check for None first, then empty
5    if arr is None:
6        print("No array provided")
7        return
8
9    if arr.size == 0:
10        print("Array is empty")
11        return
12
13    print(f"Processing {arr.size} elements")
14
15process_data(None)            # No array provided
16process_data(np.array([]))    # Array is empty
17process_data(np.array([1]))   # Processing 1 elements

Creating and Working with Empty Arrays

python

1import numpy as np
2
3# Different ways to create empty arrays
4a = np.array([])              # 1D, float64, 0 elements
5b = np.empty((0,))            # 1D, float64, 0 elements
6c = np.empty((0, 3))          # 2D, 0 rows, 3 columns
7d = np.zeros((0, 4, 5))       # 3D, 0 elements
8
9# Empty arrays preserve dtype
10e = np.array([], dtype=np.int32)
11print(e.dtype)  # int32
12
13# Concatenating with empty arrays
14arr = np.array([])
15arr = np.append(arr, [1, 2, 3])
16print(arr)  # [1. 2. 3.]
17# Warning: np.append is slow in loops — use lists and convert at the end

Common Pitfalls

Using if not array for emptiness check: NumPy arrays do not support Python truthiness for arrays with more than one element. This raises ValueError. Always use array.size == 0 instead of boolean evaluation.
Confusing np.empty() with an empty array: np.empty((3, 3)) creates a 3x3 array with uninitialized (garbage) values — it is NOT empty. np.empty((0, 3)) creates an array with zero rows, which IS empty. The name empty refers to "uninitialized", not "zero elements".
Using len() on multi-dimensional arrays: len(np.empty((3, 0))) returns 3, but the array has zero elements. len() only measures the first dimension. Use .size which returns the total element count across all dimensions.
Treating an array of zeros as empty: np.zeros(5) has 5 elements that happen to be zero. It is NOT empty (size == 5). Emptiness means no elements at all, not elements with value zero.
Building arrays with np.append in a loop: Starting with np.array([]) and appending in a loop is extremely slow (O(n^2) total time) because NumPy copies the entire array on each append. Instead, collect values in a Python list and convert once: np.array(my_list).

Summary

Use array.size == 0 to check if a NumPy array is empty — this works for all dimensions
Do not use if array: — NumPy raises ValueError for multi-element arrays
Do not use len() — it only measures the first dimension, not total element count
Check is None separately before checking .size if the variable may not be a NumPy array
np.empty((n, m)) with any zero dimension creates a truly empty array; np.empty((n, m)) with all positive dimensions creates an uninitialized (non-empty) array

How can I check whether a numpy array is empty or not?

Master System Design with Codemia

Introduction

Checking with .size

Why Not Use len()?

Why Not Use if array:?

Checking After Filtering

Checking for None vs Empty

Creating and Working with Empty Arrays

Common Pitfalls

Summary

Checking with `.size`

Why Not Use `len()`?

Why Not Use `if array:`?