Are UUID and GUID the same thing?

Yes. UUID (Universally Unique Identifier) and GUID (Globally Unique Identifier) refer to the same 128-bit identifier format. Microsoft introduced the term GUID in the 1990s, but it follows the same RFC 4122 standard that defines UUID. When you see GUID in Windows, Azure, SQL Server, or any Microsoft product, it is a UUID v4.

What does a UUID look like?

A UUID is 32 hexadecimal characters written in five groups separated by hyphens: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx. A real example looks like this: 550e8400-e29b-41d4-a716-446655440000. The total length is always 36 characters including the hyphens, or 32 characters without them.

What is the difference between UUID v4 and UUID v7?

UUID v4 is completely random. UUID v7 starts with the current Unix timestamp and fills the remaining bits with random data. The practical difference is sort order: v7 UUIDs generated later in time are always larger than earlier ones, so they insert into database indexes in order. v4 UUIDs are random and cause index fragmentation at scale. For database primary keys in new projects, v7 is the better choice. For everything else, v4 is simpler and perfectly sufficient.

Can two UUIDs ever be the same?

In theory yes, in practice no. A v4 UUID has 122 random bits, giving roughly 5.3 undecillion (5.3 x 10^36) possible values. The probability of generating the same UUID twice is so small it is not worth designing around. You would need to generate about one billion UUIDs every second for roughly 85 years before the probability of a single collision reached 50%. For any real application, treat UUID collisions as impossible.

Should I use a UUID or an auto-increment integer as a database primary key?

It depends on your needs. Auto-increment integers are simpler and smaller (4 or 8 bytes vs 16 bytes for a UUID), and they sort and index efficiently. UUIDs are better when you need to generate IDs outside the database (in your application code, across multiple databases, or in distributed systems) without a round-trip to get the next integer. UUID v7 closes much of the performance gap because it is time-sortable, making it a strong choice for distributed systems where you still want predictable insert performance.

Is a UUID the same as a hash?

No. A hash is a fixed-length fingerprint derived from input data: the same input always produces the same hash. A UUID v4 is randomly generated and has no relationship to any input. UUID v3 and v5 are exceptions: they are derived deterministically from a namespace and a name using MD5 and SHA-1 respectively, which makes them similar in concept to a hash, but they still conform to the UUID format and are not general-purpose hashes.

UUID vs GUID: What Is the Difference?

UUID and GUID are the same thing. UUID stands for Universally Unique Identifier. GUID stands for Globally Unique Identifier. Microsoft coined the term GUID in the 1990s for its own software, but the format is identical to the open UUID standard. If you see GUID in a Microsoft product and UUID everywhere else, that is why: same format, different name.

What is a UUID?

A UUID is a 128-bit number, usually written as 32 hexadecimal characters grouped by hyphens: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx. For example: 550e8400-e29b-41d4-a716-446655440000. The format is defined by the Internet Engineering Task Force (IETF) in RFC 4122. The goal is to generate identifiers that are unique across every computer, every database, and every application in the world, without any central authority assigning them.

UUID versions: v1 through v7

There are seven officially defined UUID versions. Each uses a different method to generate the identifier. The version number is the first digit of the third group (the digit after the second hyphen).

Version	How it works	Best used for
v1	Combines the current timestamp with the MAC address of the machine generating it. Produces sortable, time-ordered IDs.	Distributed systems that need time-ordered records and can tolerate leaking the generating machine's network address.
v2	Similar to v1 but embeds a POSIX user or group ID. Rarely used in practice.	Legacy enterprise systems. Avoid in new projects.
v3	Deterministic: generates a UUID from a namespace and a name using MD5 hashing. Same input always produces the same UUID.	Generating stable IDs for known resources (e.g. a UUID for a URL that must never change).
v4	Completely random. 122 of the 128 bits are randomly generated. This is the most widely used version and is what Microsoft calls a GUID.	General-purpose unique identifiers: database primary keys, API tokens, session IDs, file names.
v5	Deterministic: like v3 but uses SHA-1 instead of MD5. More collision-resistant.	Same use cases as v3, but preferred over v3 in new code because SHA-1 is stronger than MD5.
v6	A reordered version of v1, designed to sort correctly as a database index. Keeps the timestamp but reorganises the bits so the UUID sorts chronologically.	Database primary keys where insertion order matters and you want time-sortable IDs without the privacy issues of v1.
v7	Unix timestamp plus random bits. Monotonically increasing within the same millisecond. The newest version, designed for modern databases.	The best choice for new database primary keys: time-sortable, random enough to avoid collisions, and compatible with most modern databases.

Which version should you use?

For most projects, the answer is v4 or v7.

•Use v4 when you need a random, one-off identifier and do not care about sort order. It is the default in most UUID libraries and the safest choice for API tokens, session IDs, and any ID that will not be used as a database primary key.
•Use v7 when the UUID is a database primary key and you want records to insert in time order. Databases index sequential values far more efficiently than random ones, so v7 significantly reduces index fragmentation at scale.
•Use v5 (or v3) when you need the same input to always produce the same UUID. For example, generating a stable UUID for a given URL or email address.
•Avoid v1 and v2 in new projects. v1 leaks the MAC address of the generating machine, which is a minor privacy issue. v2 is practically obsolete.

How to generate a UUID

Every major programming language has a built-in or standard-library method for generating UUID v4:

javascript

// JavaScript / Node.js (v14.17+)
crypto.randomUUID()
// "3b12f1df-5232-4804-897e-917bf397618a"

// Python
import uuid
str(uuid.uuid4())
# '3b12f1df-5232-4804-897e-917bf397618a'

// PostgreSQL (generates v4 in the database)
SELECT gen_random_uuid();

If you want to generate a UUID instantly in your browser without writing any code, use AlteredIdea's UUID Generator. It runs entirely in your browser: nothing is sent to a server.

Generate a UUID right now: AlteredIdea's UUID Generator is free, browser-based, and works offline. Use the button below.

What is a UUID?

UUID versions: v1 through v7

Which version should you use?

How to generate a UUID

Related articles