How to Randomly Select from a List in Python

To randomly select from a list in Python, use random.choice(my_list) for one item. For several items without repeats, use random.sample(my_list, k); with repeats or weights, use random.choices():

import random
random.choice(cities)          # one item
random.sample(cities, 3)       # 3 different items

Each pick below is real output from Python 3.12.5 with NumPy 2.5.3. Most examples use a seed, so you can reproduce them exactly.

You’ll see choice and seeding, sample vs choices, weighted picks, removing what you picked, secure choices with secrets, and NumPy for arrays.

Pick one random item with random.choice()

choice() works on any sequence. Seed the generator when you need the same “random” result every run, like in tests:

import random

cities = ["Austin", "Denver", "Seattle", "Boston", "Miami"]
print(random.choice(cities))            # one random item

random.seed(2026)                        # same seed -> same picks every run
print([random.choice(cities) for _ in range(3)])

try:
    random.choice([])
except IndexError as err:
    print("IndexError:", err)

Output:

Boston
['Austin', 'Seattle', 'Miami']
IndexError: Cannot choose from an empty sequence

Guard against empty lists with if cities: before choosing. For random numbers rather than list items, see generating random numbers in Python.

Select several items: sample() vs choices()

sample() never repeats an item, so it can’t return more than the list holds. choices() draws with replacement, so repeats are allowed:

import random

random.seed(7)
customers = ["Emily", "James", "Olivia", "Noah", "Ava", "Liam"]

print("sample (no repeats):", random.sample(customers, 3))
print("choices (repeats ok):", random.choices(customers, k=5))

try:
    random.sample(customers, 10)
except ValueError as err:
    print("ValueError:", err)

Output:

sample (no repeats): ['Olivia', 'James', 'Noah']
choices (repeats ok): ['Noah', 'Emily', 'Noah', 'Olivia', 'Emily']
ValueError: Sample larger than population or is negative

To randomize the whole list instead, use random.shuffle(); I cover it in shuffling a list in Python.

How do you make a weighted random selection?

Pass weights to choices(). Here’s a prize draw where the big gift card should be rare, run 10,000 times to check the odds:

import random
from collections import Counter

random.seed(42)
prizes = ["$5.00 coupon", "$20.00 gift card", "$100.00 gift card"]
weights = [80, 18, 2]                    # percent chance of each prize

draws = random.choices(prizes, weights=weights, k=10_000)
for prize, n in Counter(draws).most_common():
    print(f"{prize:<18} {n:>5} draws ({n / 100:.1f}%)")

Output:

$5.00 coupon        8031 draws (80.3%)
$20.00 gift card    1765 draws (17.6%)
$100.00 gift card    204 draws (2.0%)

Weights don’t need to add up to 100; they’re relative. Building the list of results is covered in generating a list of random numbers.

Remove the picked item, use secrets, or NumPy

For a draw without putting names back, pop a random index. For codes and tokens, use secrets, which is designed for security. For big arrays, NumPy’s generator is fast and supports probabilities:

import random
import secrets

random.seed(3)
entrants = ["Emily", "James", "Olivia", "Noah"]
winner = entrants.pop(random.randrange(len(entrants)))   # pick AND remove
print("winner:", winner, "| still in the draw:", entrants)

alphabet = "ABCDEFGHJKMNPQRSTUVWXYZ23456789"
code = "".join(secrets.choice(alphabet) for _ in range(8))  # for tokens and codes
print("promo code length:", len(code), "| all from alphabet:", set(code) <= set(alphabet))

import numpy as np
rng = np.random.default_rng(1)
print(rng.choice(["small", "medium", "large"], size=6, p=[0.5, 0.3, 0.2]))

Output:

winner: James | still in the draw: ['Emily', 'Olivia', 'Noah']
promo code length: 8 | all from alphabet: True
['medium' 'large' 'small' 'large' 'small' 'small']

Never use random for passwords or reset links; its output can be predicted. NumPy’s random tools are covered in NumPy random in Python.

Which random selection method should you use?

I start with choice() or sample() and only reach for weights, secrets or NumPy when the job needs them. Removing items afterwards is covered in removing an element from a list by index.

Goal	Use
One item	random.choice(lst)
Several unique items	random.sample(lst, k)
Several items, repeats or weights	random.choices(lst, weights=..., k=n)
Pick and remove	lst.pop(random.randrange(len(lst)))
Security-sensitive	secrets.choice(lst)
Large arrays	numpy.random.default_rng().choice()

These guides go well with random selection:

• Generate a random string of a specific length
• Generate random numbers between 0 and 1
• Find duplicates in a Python list
• Remove the last element from a list

Frequently asked questions