UUID Data Type

25.12.1UUIDData Type

Overview

MonkDB provides a first-class UUID type for table columns and expressions.

• Can be used as primary key or regular column.
• Stored/indexed as exact-match keyword-like field.
• Type preserved in mappings using metadata (monkdb_type).
• PostgreSQL OIDs supported (2950 for uuid, 2951 for uuid[]).

Basic usage

CREATE TABLE doc.users (
  id UUID PRIMARY KEY,
  name TEXT,
  created_at TIMESTAMPTZ
);

CREATE TABLE doc.orders (
  id UUID PRIMARY KEY DEFAULT gen_random_text_uuid(),
  amount DOUBLE PRECISION
);

SELECT *
FROM doc.orders
WHERE id = '550e8400-e29b-41d4-a716-446655440000';

UUID generation functions

• gen_random_uuid()
• uuid_generate_v4() (alias)
• gen_random_text_uuid()

Casting and validation behavior

Accepted inputs include:

• Canonical UUID string (xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx)
• UUID values
• BytesRef/string-like values that can be interpreted as UUID text
• URL-safe/base64 UUID bytes and text-UUID compatibility paths

Invalid UUID values are rejected by type casting/validation.

Indexing guidance

Do:

• Use UUID as PK for entity identity.
• Index additional columns used in filters.

Don't:

• Add redundant index on UUID PK column.
• Over-index composite keys that always begin with UUID unless query patterns require it.