Schemas & Types

Data Types

Type	Description	Example
`STRING`	UTF-8 text	`'hello'`
`TEXT`	Alias for STRING	`'hello'`
`VARCHAR`	Alias for STRING	`'hello'`
`INT`	64-bit signed integer	`42`
`FLOAT`	64-bit IEEE 754	`3.14`
`DECIMAL`	Arbitrary precision	`99.99`
`BOOL`	Boolean	`true`
`TIMESTAMP`	UTC timestamp (nanosecond precision)	`now()`
`DATETIME`	Alias for TIMESTAMP	`'2026-01-01T00:00Z'`
`UUID`	128-bit UUID	`gen_uuid_v7()`
`GEOMETRY`	OGC geometry (WKB internally)	`ST_Point(-73.9, 40.7)`
`ARRAY`	Ordered list	`ARRAY[1, 2, 3]`
`OBJECT`	Nested document	`{ name: 'Alice' }`

Schema Modes

Schemaless — No schema required. Fields can vary between documents. Types are inferred on insert. This is the default when you CREATE COLLECTION x.

Strict — Schema is defined at creation time and enforced on every write. O(1) field extraction via binary tuple format. Created with TYPE DOCUMENT STRICT (...).

Typeguards (Schemaless Validation)

Typeguards add write-time validation to schemaless collections without changing the storage format. Guarded fields are type-checked; unguarded fields pass freely.

CREATE TYPEGUARD ON users (
    email STRING REQUIRED CHECK (email LIKE '%@%.%'),
    age INT CHECK (age >= 0 AND age <= 150),
    role STRING DEFAULT 'user',
    updated_at TIMESTAMP VALUE now()
);

Modifier	Behavior
`REQUIRED`	Field must be present and non-null
`DEFAULT`	Inject a value when the field is absent
`VALUE`	Always inject/overwrite (computed fields)
`CHECK`	SQL boolean expression validated at write time

-- Modify guards
ALTER TYPEGUARD ON users ADD score FLOAT CHECK (score >= 0);
ALTER TYPEGUARD ON users DROP age;

-- Audit existing data against guards
VALIDATE TYPEGUARD ON users;

-- Graduate to strict schema
CONVERT COLLECTION users TO strict;

Schema Evolution (Strict Mode)

-- Add a column with a default (zero-downtime, multi-version reads)
ALTER COLLECTION orders ADD COLUMN region STRING DEFAULT 'us-east';

-- Drop a column
ALTER COLLECTION orders DROP COLUMN region;

ALTER ADD COLUMN uses multi-version reads — existing rows are served with the default value without rewriting storage. New writes include the column natively.