UUID / ULID Generator: Frequently Asked Questions

UUID v4 has 122 random bits, giving 2^122 ≈ 5.3×10^36 possible values. The probability of generating a duplicate among 1 trillion UUIDs is approximately 1 in 5.3×10^24 — effectively zero for any practical application. They are considered unique for all practical purposes.

UUID (Universally Unique Identifier) and GUID (Globally Unique Identifier) refer to the same concept. UUID is the RFC 4122 standard term. GUID is Microsoft's terminology. They use the same format and are interchangeable.

Yes, but be aware of the index fragmentation issue with UUID v4 in SQL databases. For best performance, use time-ordered identifiers (ULID, UUID v7) or use UUID as a secondary unique column with an integer primary key.

UUID v4 is the safe default for most applications. Use UUID v1 if you need embedded timestamps for debugging. Use ULID for time-sortable records and better database index performance. The older UUID v3 and v5 (name-based, deterministic) are useful for generating consistent IDs from stable inputs.

UUID values are hexadecimal and case-insensitive — 6ba7b810-9dad-11d1-80b4-00c04fd430c8 and 6BA7B810-9DAD-11D1-80B4-00C04FD430C8 are the same UUID. Most systems store and display them in lowercase by convention.

In JavaScript: crypto.randomUUID() (built-in since Node 14.17 and all modern browsers). In Python: import uuid; str(uuid.uuid4()). In PostgreSQL: gen_random_uuid(). In MySQL: UUID(). In shell: uuidgen (macOS/Linux).