Hash Function

A function mapping data of any size to a fixed-size value.

Backend Engineer · Junior Frontend Engineer · Junior Data Engineer · Junior

Also known as: hashing, hash, checksum

A hash function takes input of any size (a string, a file, a record) and produces a fixed-size value, called a hash or digest. The same input always gives the same output, and a tiny change in the input produces a very different output.

import hashlib
hashlib.sha256(b"hello").hexdigest()
# '2cf24dba5fb0a30e26e83b2ac5b9e29e1b161e5c1fa7425e73043362938b9824'
hashlib.sha256(b"hellp").hexdigest()
# a completely different value

Properties

Deterministic: same input, same hash.
Fast to compute.
Fixed-size output, however big the input.
Spreads values evenly, so different inputs rarely share a hash (a collision).
For cryptographic hashes: one-way (you can’t recover the input from the hash) and collision-resistant.

Two families, with different jobs

Kind	Examples	Used for
Non-cryptographic	the hash behind dictionaries, MurmurHash, xxHash	hash tables, partitioning, quick checks
Cryptographic	SHA-256, SHA-3	integrity checks, signatures, content addressing
Password hashes	Argon2, bcrypt	deliberately slow, for storing passwords (password hashing)

Common uses

Dictionaries and sets: the hash picks where to store a key.
Integrity: compare a file’s hash to a published one to detect corruption or tampering.
Deduplication and caching: identify identical content or cache keys.
Sharding: choose a partition from hash(key) % N.
Git names objects by their hash.

Cautions

Hashing is not encryption. You can’t decrypt a hash (hashing vs encryption).
Don’t use fast general-purpose hashes (MD5, SHA-256) for passwords, and don’t use broken ones (MD5, SHA-1) for security.
Don’t invent your own.
Python randomizes string hashes between runs, so don’t store hash(x) or depend on its value.