Fix "ValueError: operands could not be broadcast" in NumPy

When performing element-wise arithmetic operations or matrix operations in NumPy, you may encounter the following error:

ValueError: operands could not be broadcast together with shapes (3,4) (3,)

Why Does This Error Occur?

NumPy uses broadcasting rules to perform element-wise operations on arrays of different shapes. For broadcasting to work, trailing dimensions must either:

Be equal in size. One of the dimensions must be 1. If neither condition is met, NumPy cannot align the arrays and raises this error.

Minimal Reproducible Example

The error occurs when attempting to add a 1D array across columns without matching dimensions:

import numpy as np

# Shape: (3, 4)
matrix = np.ones((3, 4))

# Shape: (3,)
vector = np.array([1, 2, 3])

# Fails with ValueError
result = matrix + vector

NumPy tries to match the trailing dimension 4 with 3, which fails.

Solution 1: Use np.newaxis to Reshape for Column Broadcasting

If your goal is to add the vector to each column, add an axis to make the shape (3, 1):

import numpy as np

matrix = np.ones((3, 4))
vector = np.array([1, 2, 3])

# Reshape vector to (3, 1)
result = matrix + vector[:, np.newaxis]

print(result.shape)  # Output: (3, 4)

Solution 2: Reshape Using .reshape(-1, 1)

Alternatively, use the .reshape() method:

result = matrix + vector.reshape(-1, 1)

Both methods make the array compatible for broadcasting according to NumPy's dimension alignment rules.