Python:NumPy .all()

In NumPy, the .all() method returns True if all elements in an ndarray evaluate to True, or if all elements along a specified axis evaluate to True.

The .all() method can operate on the entire array to check if all values are truthy, or work along specific axes to perform row-wise or column-wise boolean validation. It follows Python’s truth value testing conventions where non-zero numbers, non-empty arrays, and True values are considered truthy, while zero, empty arrays, None, and False are considered falsy. This method is particularly useful in data validation, filtering operations, and quality control checks in data science workflows.

Syntax

ndarray.all(axis=None, out=None, keepdims=False, where=True)

Parameters:

• axis (optional): Specifies the axis or axes along which to perform the logical AND operation.
  • If None (default), the test is applied to all elements in the array.
  • If integer, checks along that specific axis.
  • If tuple of integers, checks along multiple axes.
• out (optional): Alternative output array to store the result. Must have the same shape as the expected output.
• keepdims (optional): If True, the reduced axes are retained in the result as dimensions with size one. If False (default), the axes are removed from the result.
• where (optional): Boolean array indicating which elements to include in the check. Elements where the condition is False are ignored and assumed to be True.

Return value:

Returns a boolean or ndarray of booleans, depending on the axis argument:

• A single bool if the check is performed on the entire array.
• An ndarray of bool values if performed along a specific axis.

Example 1: Basic Boolean Validation

This example demonstrates using .all() to validate boolean conditions across an array:

import numpy as np

# Create arrays with different boolean patterns
all_true = np.array([True, True, True, True])
has_false = np.array([True, True, False, True])
all_positive = np.array([1, 5, 3, 7])
has_zero = np.array([1, 5, 0, 7])

print("All True array:", all_true.all())
print("Array with False:", has_false.all())
print("All positive numbers:", all_positive.all())
print("Array with zero:", has_zero.all())

# Using comparison operators
data = np.array([85, 92, 78, 95, 88])
all_passing = (data >= 70).all()
all_excellent = (data >= 90).all()

print(f"\nAll scores >= 70: {all_passing}")
print(f"All scores >= 90: {all_excellent}")

Copy to clipboard
Copy to clipboard

The output of this code is:

All True array: True
Array with False: False
All positive numbers: True
Array with zero: False

All scores >= 70: True
All scores >= 90: False

Copy to clipboard
Copy to clipboard

This example shows how .all() evaluates different types of arrays. Non-zero numbers are considered truthy, while zero and False values cause the method to return False.

Example 2: Axis-wise Validation in Multi-dimensional Arrays

This example demonstrates how to validate conditions along specific axes in a 2D array:

import numpy as np

# Create a 2D array of test results (pass=1, fail=0)
test_results = np.array([
    [1, 1, 1, 1],  # Student 1: All passed
    [1, 1, 0, 1],  # Student 2: One failure
    [1, 1, 1, 1],  # Student 3: All passed
    [1, 0, 1, 1]   # Student 4: One failure
])

print("Test results (1=pass, 0=fail):")
print(test_results)

# Check if all students passed each test (column-wise)
all_passed_per_test = test_results.all(axis=0)
print(f"\nAll students passed each test: {all_passed_per_test}")

# Check if each student passed all tests (row-wise)
all_passed_per_student = test_results.all(axis=1)
print(f"Each student passed all tests: {all_passed_per_student}")

# Check if all students passed all tests (entire array)
perfect_class = test_results.all()
print(f"Perfect class (all passed): {perfect_class}")

Copy to clipboard
Copy to clipboard

The output of this code is:

Test results (1=pass, 0=fail):
[[1 1 1 1]
 [1 1 0 1]
 [1 1 1 1]
 [1 0 1 1]]

All students passed each test: [ True False False  True]
Each student passed all tests: [ True False  True False]
Perfect class (all passed): False

Copy to clipboard
Copy to clipboard

This example shows how the axis parameter controls the direction of validation. Using axis=0 checks columns (tests), while axis=1 checks rows (students).

Example 3: Data Quality Validation with keepdims

This example shows how to use .all() with keepdims for data quality checks that maintain array dimensions:

import numpy as np

# Create sensor data with some readings potentially out of range
sensor_data = np.array([
  [45, 52, 48, 51],
  [46, 49, 47, 50],
  [44, 53, 46, 49]
])

print("Sensor readings:")
print(sensor_data)

# Define valid range (40-55)
min_valid = 40
max_valid = 55

# Check if all readings per sensor are within range
within_range = (sensor_data >= min_valid) & (sensor_data <= max_valid)
all_valid_per_sensor = within_range.all(axis=1, keepdims=True)

print(f"\nAll readings valid per sensor:\n{all_valid_per_sensor}")
print(f"Shape: {all_valid_per_sensor.shape}")

# Create a quality report
quality_report = np.where(all_valid_per_sensor, "PASS", "FAIL")
print(f"\nQuality report:\n{quality_report}")

Copy to clipboard
Copy to clipboard

The output of this code is:

Sensor readings:
[[45 52 48 51]
 [46 49 47 50]
 [44 53 46 49]]

All readings valid per sensor:
[[ True]
 [ True]
 [ True]]
Shape: (3, 1)

Quality report:
[['PASS']
 ['PASS']
 ['PASS']]

Copy to clipboard
Copy to clipboard

The keepdims=True parameter maintains array dimensions, making it easier to combine validation results with other operations or create aligned reports.

Codebyte Example: Using the where Parameter for Conditional Validation

This example demonstrates selective validation using the where parameter:

Visit us
Visit us
Hide code
Code
Output
Hide output
Loading...
Copy to your clipboard
Copy Codebyte
Run

The where parameter allows selective validation, ignoring certain elements (like discontinued products) when checking conditions.