pandas and NumPy 3 cheatsheet

pandas and NumPy 3: NumPy

10 common pandas and NumPy 3 tasks.

Create an array

Create an array from a list. Arrays hold one type and operate on all items at once.

a = np.array([1, 2, 3])

Create evenly spaced values, either by step or by count.

np.arange(0, 10, 2)
np.linspace(0, 1, 5)

Create arrays of zeros, ones or a given value.

np.zeros((3, 4))
np.ones(5)
np.full((2, 2), 7)

Change the shape without changing the data. -1 lets NumPy work out the size.

a.reshape(2, -1)

Use the modern generator with a seed so results can be reproduced.

rng = np.random.default_rng(42)
rng.normal(size=5)
rng.integers(0, 10, size=3)

Average each column (axis 0) or each row (axis 1).

a.mean(axis=0)
a.std(axis=1)
a.sum()

Select the items that satisfy a condition.

a[a > 2]

Use the @ operator for matrix multiplication.

c = a @ b

Sort values, or get the indices that would sort them.

np.sort(a)
np.argsort(a)

Get the sorted unique values and how often each occurs.

values, counts = np.unique(a, return_counts=True)