larzuuid

Modern, sortable IDs — UUIDv7, ULID, nanoid. Pure Python, zero dependencies.

Random UUIDv4s are unguessable but scatter across a database index, hurting insert performance and locality. Time-ordered ids fix that: they sort by creation time, cluster in the index, and let you read the timestamp back out. The stdlib uuid module has none of these — larzuuid adds the three you actually want.

from larzuuid import uuid7, ulid, nanoid

uuid7()     # 018f5e...-7...  a real UUID, but time-sortable
ulid()      # 01HZY3K5...     26-char Crockford base32, sortable
nanoid()    # V1StGXR8_Z5jdHi6B-myT   compact, URL-safe

Why

• UUIDv7 (RFC 9562) — a standard uuid.UUID, but the first 48 bits are a millisecond timestamp, so it sorts by time and works as a great primary key. Read the time back with uuid7_timestamp().
• ULID — 26 lexicographically-sortable Crockford-base32 chars; MonotonicULID guarantees strictly increasing ids even within the same millisecond.
• nanoid — a compact, URL-safe random id (21 chars by default), uniform via rejection sampling, with a custom alphabet option.
• Zero dependencies. No ulid-py, no nanoid, no uuid6.

Install

pip install larzuuid

Usage

from larzuuid import uuid7, uuid7_timestamp, ulid, ulid_timestamp, MonotonicULID, nanoid

pk = uuid7()                       # use as a DB key; sorts by creation time
uuid7_timestamp(pk)                # datetime (UTC)

u = ulid(); ulid_timestamp(u)
gen = MonotonicULID(); gen.new(); gen.new()   # strictly increasing

nanoid()          # 21 chars
nanoid(10, alphabet="0123456789abcdef")

Tests

python -m unittest discover -s tests -v   # 13 tests

The Larz stack

One of 30+ pure-Python, zero-dependency libraries at github.com/larz-scripter.

License

MIT © larz-scripter