Developer tooling around UUIDs moves faster than the standards themselves. PostgreSQL 18 lands native UUID v7 support. Language runtimes add generation to the standard library. ORMs add type-aware columns. This section tracks those changes with concrete detail: what shipped, which version you need, and whether it requires any action in your project.
What we cover
Articles here focus on changes that affect how developers generate, store, and query UUIDs in real applications. A new native function in a database release gets covered if it changes the recommended approach. A library release gets covered if it adds v7 support or changes a default. We skip changelog noise and focus on engineering decisions.
How to use this as a reference
Each article stands alone as a technical reference. If you arrived from a search about PostgreSQL 18 and UUID v7, the article answers the specific question — what the function is called, how to use it, and what the migration path looks like — without assuming you have read anything else here.
If a topic here leads you to want the full background, the Guides section covers UUID versions, database storage, security, and language-specific implementation in depth.
Latest articles
UUID in SQLite — TEXT vs BLOB and WITHOUT ROWID
SQLite has no native UUID column type. TEXT wastes space; BLOB with WITHOUT ROWID matches B-tree performance. Here are the storage options and UUID v7 patterns for Python, Go, and Node.js.
-
UUID URL-Safe Encoding — Base64url, Base58, and NanoID
Standard UUID strings are 36 characters. Base64url shrinks to 22, Base58 to 22 with no padding. Here is when shorter encodings make sense and how to encode and decode correctly.
-
UUID v7 and Clock Skew — What Happens When Clocks Drift
UUID v7 relies on the system clock. What happens when NTP adjusts time backward, a VM is migrated, or a container clock drifts — and how RFC 9562 §6.2 mitigates these failure modes.
-
UUID as Idempotency Keys — Patterns and Best Practices
Idempotency keys prevent duplicate charges and double-submits. UUID v4 is the correct version here — v7's timestamp makes it guessable. Here is the full pattern and storage strategies.
-
UUID in .NET and C# — Guid, v7, and EF Core
.NET 9 adds Guid.CreateVersion7() to the standard library. Here is how to use it with Entity Framework Core, handle SQL Server's byte-order caveat, and configure Npgsql for PostgreSQL.
-
UUID in PHP and Laravel — v4, v7, and Eloquent
Laravel's HasUuids trait defaults to UUID v4. One method override switches to UUID v7. Here is the full pattern with ramsey/uuid, Eloquent, and MySQL binary(16) storage.
-
UUID in Rust — uuid Crate, v7 Support, and Diesel
The uuid crate's Uuid::now_v7() gives you RFC 9562-compliant time-ordered IDs in Rust. Here is how to wire it into Diesel, SQLx, and serde for a production-ready setup.
-
UUID v7 vs ULID — Which Should You Use?
Both are time-sortable 128-bit IDs. UUID v7 is the RFC standard with PostgreSQL 18 support; ULID has a shorter string and more random bits. Here is how to decide.
-
UUID v8 — Custom Layout for Application-Specific IDs
UUID v8 gives you 122 free bits to define your own layout — embed shard IDs, custom timestamps, or region codes while staying RFC 9562 compliant.
-
UUID in Java and Spring Boot — v4, v7, and JPA
Hibernate 6.5+ generates UUID v7 with a single annotation. Here is how to configure JPA entity IDs, Diesel-style integration, and BINARY(16) for MySQL backends.
-
UUID in React and Next.js — Patterns and Pitfalls
Generating UUIDs in React components causes hydration mismatches. Here is where to generate them, how to use them as stable keys, and the Next.js App Router patterns.
-
TypeScript UUID Types — Branded Types and Validation
Plain string UUIDs let you pass a UserId where an OrderId is expected. Branded types catch ID mixups at compile time with zero runtime cost — here is the full pattern.
-
Migrating from UUID v4 to v7 in Production
You do not need to touch existing v4 rows. Here is the zero-downtime path — PostgreSQL, MySQL, Node.js, Python, Go, Prisma, Drizzle, and Django — with a safe rollback.
-
UUID in MongoDB — Replacing ObjectId with UUID v7
ObjectId is MongoDB-specific. UUID v7 gives you the same time-ordering with RFC standardisation and cross-database portability. Here is how to switch and what to expect.
-
MySQL UUID Storage — BINARY(16) vs VARCHAR(36)
VARCHAR(36) is the most common MySQL UUID mistake — it doubles storage and slows indexes. Here is why BINARY(16) is correct and how to migrate an existing table.
-
UUID v7 in Distributed Systems
UUID v7's millisecond timestamp enables event ordering and log correlation without a central sequence counter. Here is how to use it across multi-node architectures.
-
PostgreSQL 18 Ships Native UUID v7
PostgreSQL 18 adds a built-in uuidv7() function, making time-sortable primary keys a single SQL call away. Here is what it means for your schema and migration path.
-
crypto.randomUUID() — Browser and Runtime Support
crypto.randomUUID() is now universally available — Chrome 92+, Firefox 95+, Safari 15.4+, Node.js, Deno, and Bun. Here is the full matrix and when you still need a library.
-
UUID Primary Keys vs Auto-Increment — 2026 Guide
Auto-increment is simpler, UUID is right for distributed systems. UUID v7 closes the performance gap — here is the full trade-off and when each is correct.
-
RFC 9562 Replaces RFC 4122 — What Changed
The current UUID standard adds v6, v7, v8, and the Max UUID sentinel. It also formally deprecates v1 and v2. Here is every change and what it means in practice.
-
UUID v7 ORM Support in 2026
All four major ORMs now support UUID v7 natively. Here is the current state of each and how to configure application-layer vs database-level generation.
-
Python 3.13 Adds uuid.uuid7() to the Standard Library
No third-party package needed on Python 3.13+. Here is how uuid.uuid7() works, how to migrate from the uuid6 backport, and patterns for Django, SQLAlchemy, and FastAPI.
Frequently asked questions
Does PostgreSQL 18 support UUID v7 natively?
Yes. PostgreSQL 18 ships a built-in uuidv7() function that generates RFC 9562-compliant UUID v7 values directly in SQL — no extensions or application-level generation required. The values store in the existing uuid column type unchanged.
Why is UUID v7 becoming the default recommendation in 2026?
UUID v7 combines the global uniqueness of UUIDs with a millisecond-precision timestamp prefix, making inserts sequential in B-tree indexes and eliminating the page-split fragmentation that v4 causes at scale. RFC 9562 standardised it in 2024, and PostgreSQL 18 adding native support in 2026 makes it the obvious default for new projects.
Should I use v4 or v7?
Use v7 for database primary keys (time-sortable, index-friendly) and v4 for anything where creation order could leak information, like tokens or share links.
What is the latest UUID standard?
The current standard is RFC 9562, published by the IETF in May 2024. It supersedes RFC 4122 (2005), deprecates UUID v1 and v2, and formally introduces UUID v7 as the recommended format for database primary keys. The Wikipedia UUID article maintains a current summary of the standard.
Which databases now support UUID v7 natively?
PostgreSQL 18 ships a built-in uuidv7() function. MySQL and SQLite do not have native v7 support but accept UUID v7 strings in standard UUID columns. Python 3.13 added uuid.uuid7() to the standard library. ORMs including Prisma, Drizzle, and SQLAlchemy have added v7 configuration options in their recent releases.
How do I stay up to date with UUID-related changes?
Follow the IETF RFC search for UUID for standard updates, and watch the ietf-wg-uuidrev GitHub repository for any future revisions. The UUID News section on this site tracks database releases, library updates, and standard changes with practical developer summaries.