Reshaping

Code

import numpy as np

a = np.arange(12)

# Reshape and -1 inference
b = a.reshape(3, 4)
c = a.reshape(-1, 2)        # 6 rows inferred
flat = b.flatten()          # copy
ravel = b.ravel()           # view when possible

# Transpose and swap axes
t = b.T
swapped = b.swapaxes(0, 1)

# Stacking
x = np.array([1, 2, 3])
y = np.array([4, 5, 6])
vstack = np.vstack([x, y])     # rows
hstack = np.hstack([x, y])     # columns
concat = np.concatenate([x, y])

print(b.shape, c.shape, vstack.shape)

Explanation

reshape returns a view with the same data in a new shape, and -1 lets NumPy infer one dimension automatically. flatten copies while ravel usually returns a view, so prefer ravel when mutating. vstack and hstack concatenate along new and existing axes respectively.

More NumPy Snippets