Python ULID

Release v4.2.1.dev3+g4e4c24f47 (What’s new)

A ULID is a universally unique lexicographically sortable identifier. It is

  • 128-bit compatible with UUID

  • 1.21e+24 unique ULIDs per millisecond

  • Lexicographically sortable!

  • Canonically encoded as a 26 character string, as opposed to the 36 character UUID

  • Uses Crockford’s base32 for better efficiency and readability (5 bits per character)

  • Case insensitive

  • No special characters (URL safe)

  • Monotonic sort order (correctly detects and handles the same millisecond)

In general the structure of a ULID is as follows:

 01AN4Z07BY      79KA1307SR9X4MV3
|----------|    |----------------|
 Timestamp          Randomness
   48bits             80bits

For more information have a look at the original specification.

Installation

Use pip to install the library

$ pip install python-ulid

to include Pydantic support install the optional dependency like so

$ pip install python-ulid[pydantic]

and for the SQLAlchemy column type

$ pip install python-ulid[sqlalchemy]

Basic Usage

Create a new ULID object from the current timestamp

>>> from ulid import ULID
>>> ULID()
ULID(01E75HZVW36EAZKMF1W7XNMSB4)

or use one of the named constructors

>>> import time, datetime
>>> ULID.from_timestamp(time.time())
ULID(01E75J1MKKWMGG0N5MBHFMRC84)
>>> ULID.from_datetime(datetime.datetime.now())
ULID(01E75J2XBK390V2XRH44EHC10X)

There are several options for encoding the ULID object (e.g. string, hex, int, bytes, UUID):

>>> str(ulid)
'01BTGNYV6HRNK8K8VKZASZCFPE'
>>> ulid.hex
'015ea15f6cd1c56689a373fab3f63ece'
>>> int(ulid)
1820576928786795198723644692628913870
>>> bytes(ulid)
b'\x01^\xa1_l\xd1\xc5f\x89\xa3s\xfa\xb3\xf6>\xce'
>>> ulid.to_uuid()
UUID('015ea15f-6cd1-c566-89a3-73fab3f63ece')

It is also possible to directly access the timestamp component of a ULID, either in UNIX epoch or as datetime.datetime

>>> ulid.timestamp
1505945939.153
>>> ulid.datetime
datetime.datetime(2017, 9, 20, 22, 18, 59, 153000, tzinfo=datetime.timezone.utc)

Pydantic integration

The ULID class can be directly used for the popular data validation library Pydantic like so

from pydantic import BaseModel
from ulid import ULID


class Model(BaseModel):
  ulid: ULID

model = Model(ulid="DX89370400440532013000")  # OK
model = Model(ulid="not-a-ulid")  # Raises ValidationError

SQLAlchemy and SQLModel integration

The ULIDType column type stores ULID values with SQLAlchemy (version 2.0 or newer). Values are bound from ULID, UUID, bytes, int or any string representation and always loaded as ULID. Pass default=ULID to generate a new identifier for every inserted row

from sqlalchemy.orm import DeclarativeBase, Mapped, mapped_column
from ulid import ULID
from ulid.sqlalchemy import ULIDType


class Base(DeclarativeBase):
  pass


class Item(Base):
  __tablename__ = "items"

  id: Mapped[ULID] = mapped_column(ULIDType(), primary_key=True, default=ULID)
  name: Mapped[str]

By default the value is stored as a UUID: a native UUID column where the database has one (e.g. PostgreSQL) and CHAR(32) otherwise. Use ULIDType(storage="binary") for a BINARY(16) column or ULIDType(storage="string") for the 26 character ULID string in a CHAR(26) column. All storages keep the database sort order identical to the ULID order, except the native UNIQUEIDENTIFIER type on SQL Server.

SQLModel models can use ULID as a field type directly, since it is a Pydantic type. Only the column type has to be passed explicitly

from sqlmodel import Field, SQLModel
from ulid import ULID
from ulid.sqlalchemy import ULIDType


class Hero(SQLModel, table=True):
  id: ULID = Field(default_factory=ULID, primary_key=True, sa_type=ULIDType)
  name: str

Monotonic Support

This library by default supports the implementation for monotonic sort order suggested by the official ULID specification.

This means that ULID values generated in the same millisecond will have linear increasing randomness values. If \(r_1\) and \(r_2\) are the randomness values of two ULIDs with the same timestamp, then \(r_2 = r_1 + 1\).

Generators and policies

Every ULID is produced by a ULIDGenerator, which samples a clock for the timestamp, sources entropy for the randomness, and enforces a monotonicity policy. The bare ULID() constructor and the ULID.from_* factory methods delegate to a shared module-level default_generator.

For most use cases the default is all you need. To customize generation — a different clock, a custom entropy source, or another monotonicity policy — create your own ULIDGenerator and call generate()

>>> from ulid import ULIDGenerator
>>> generator = ULIDGenerator()
>>> generator.generate()
ULID(01HB0N8Q4RCE7YB1M2VZK9WX3T)

Choosing a monotonicity policy

A policy decides how the randomness is resolved when multiple ULIDs are generated within the same millisecond. Three policies are available:

  • StrictMonotonicPolicy (default) — increments the randomness by 1 on a same-millisecond collision and raises ValueError if the randomness is exhausted. This is the behaviour described under Monotonic Support.

  • LaxMonotonicPolicy — increments like the strict policy, but regenerates fresh randomness instead of raising when the randomness is exhausted.

  • PureRandomPolicy — ignores previous state and always draws fresh randomness, maximizing entropy at the cost of same-millisecond sort order.

>>> from ulid import ULIDGenerator, PureRandomPolicy
>>> generator = ULIDGenerator(policy=PureRandomPolicy())
>>> generator.generate()
ULID(01HB0N9F3TA5KDQ6ZE0WYV7MRC)

Overriding the default generator

To make ULID() and the ULID.from_* constructors use a custom generator globally, reassign ulid.default_generator

>>> import ulid
>>> from ulid import ULID, ULIDGenerator, LaxMonotonicPolicy
>>> ulid.default_generator = ULIDGenerator(policy=LaxMonotonicPolicy())
>>> ULID()  # now generated with the lax policy
ULID(01HB0NB7X2M4C8VKQ0ZF5WD9RA)

Command line interface

The package comes with a CLI interface that can be invoked either by the script name ulid or as python module python -m ulid. The CLI allows you to generate, inspect and convert ULIDs, e.g.

$ ulid build
01HASFKBN8SKZTSVVS03K5AMMS

$ ulid build --from-datetime=2023-09-23T10:20:30
01HB0J0F5GCKEXNSWVAD5PEAC1

$ ulid show 01HASFKBN8SKZTSVVS03K5AMMS
ULID:      01HASFKBN8SKZTSVVS03K5AMMS
Hex:       018ab2f9aea8ccffacef7900e6555299
Int:       2049395013039097460549394558635823769
Timestamp: 1695219822.248
Datetime:  2023-09-20 14:23:42.248000+00:00

There are several flags to select specific output formats for the show command, e.g.

$ ulid show --datetime 01HASFKBN8SKZTSVVS03K5AMMS
2023-09-20 14:23:42.248000+00:00

The special character - allows to read values from stdin so that they can be piped. E.g.

$ echo 01HASFKBN8SKZTSVVS03K5AMMS | ulid show --uuid -
018ab2f9-aea8-4cff-acef-7900e6555299

$ date --iso-8601 | python -m ulid build --from-datetime -
01HAT9PVR02T3S13XB48S7GEHE

For a full overview of flags for the build and show commands use the --help option (e.g. ulid show --help).

API documentation