Question 1

how does ulid differ from uuid

Accepted Answer

A ULID is 26 Crockford Base32 characters, the first ten of which are time in milliseconds, so values sort by creation. UUID v4 is 36 characters in random order; UUID v7 is built like ULID but written in hexadecimal and standardised in RFC 9562. Both formats take 128 bits; they differ in notation and in having a standard.

Question 2

how to get the time from a ulid

Accepted Answer

The first ten characters are 48 bits of time in Crockford Base32: five bits per character, high bits first. Convert them to milliseconds since 1 January 1970 to get the creation moment: 01ARYZ6S41 is 30 July 2016, 22:36:16 UTC. The other sixteen characters are the random part and hold no time.

Question 3

why does ulid have no i l o u

Accepted Answer

That is the Crockford Base32 alphabet: it drops I and L, confused with one, O, confused with zero, and U so that rude words do not form by accident. When decoding, many implementations accept I and L as 1, O as 0, and ignore case. That makes a ULID easy to dictate and copy by hand.

Question 4

can a ulid be stored in a uuid column

Accepted Answer

As a binary value, yes: both ULID and UUID take 128 bits, and many libraries convert one into the other. But the result is a UUID without proper version and variant bits, and strict checks will call it non-standard. A 26-character string will not fit a uuid column. If you need the standard type and ordering, use UUID v7 to begin with.

Question 5

do ulids from the same millisecond sort

Accepted Answer

If the random part is fresh each time, no — order within a millisecond is random. The ULID spec describes a monotonic mode: within the same millisecond the random part is incremented by one. Our generator works exactly that way, so a batch of fifty values issued within one millisecond comes out strictly ascending.

Question 6

until what year will ulid last

Accepted Answer

Time in a ULID is 48 bits of milliseconds since 1970, enough until the year 10889, as the spec itself says. That is why the first character of a ULID is never above 7 — the top bits are still zero. A string starting with 8 or higher is not a ULID, and time decoding rejects it.

Format	Length	Randomness	Time inside	Sortable	Standard
UUID v4	36	122 bits	no	no	RFC 9562
UUID v7	36	74 bits	ms, 48 bits	yes	RFC 9562
UUID v1	36	node and counter	100 ns, 60 bits	no	RFC 9562, legacy
Nil UUID	36	none	no	—	RFC 9562
NanoID	21	126 bits	no	no	a library, not an RFC
ULID	26	80 bits	ms, 48 bits	yes	ULID spec, not an RFC

UUID generator · formatULID — a sortable identifier

Generator ULID

How it is built: ULID

When to use it

How it differs from its neighbours

Format comparison

Frequently asked questions

how does ulid differ from uuid

how to get the time from a ulid

why does ulid have no i l o u

can a ulid be stored in a uuid column

do ulids from the same millisecond sort

until what year will ulid last

UUID generator · formatULID — a sortable identifier

Generator ULID

How it is built: ULID

When to use it

How it differs from its neighbours

Format comparison

Frequently asked questions

how does ulid differ from uuid

how to get the time from a ulid

why does ulid have no i l o u

can a ulid be stored in a uuid column

do ulids from the same millisecond sort

until what year will ulid last

Related tools

Password generator

Random numbers

Unix Timestamp

Hash functions