Schema Definition
an5 uses a declarative schema syntax to define your data models. Schema files use the .an5 extension.
Basic Syntax
model User {
id NVARCHAR(1000) @id @default(uuid())
email NVARCHAR(255) @unique
name NVARCHAR(255)?
createdAt DATETIME2 @default(now())
@@map("users")
}
Model Definition
Fields
Each field has a name, type, and optional attributes:
model Post {
id NVARCHAR(1000) @id @default(uuid())
title NVARCHAR(255) @description("The post title")
content TEXT?
published BIT @default(false)
authorId NVARCHAR(1000)
}
Active Context: Targeting Sqlite with Golang client generator.
Ensure field types in your .an5 models align with the Sqlite dialect requirements below.
Field Types
Valid types depend on the database you are generating for. The provider is read from
the connection string (sqlserver://, postgres://, mysql://, sqlite://,
googlesheets://, or a path ending in .sqlite/.sqlite3/.db); with no connection
string it is SQL Server. A type the provider does not have stops generation with the
field that used it, instead of producing SQL the database rejects later.
This is stricter than before, and deliberately: one shared list used to accept
INTEGER for SQL Server and BIT for SQLite, so a schema only failed once the
database rejected it. Migrating means replacing each type with the provider’s own —
BOOLEAN → BIT on SQL Server, DATETIME2 → DATETIME or DATETIME2 kept on SQL
Server but DATETIME on SQLite, and so on. The error names the provider and the
field, so a run over the schema lists everything to change at once.
Names that would read as a relation are never types: USER and NAME are real
PostgreSQL types but are left out on purpose, so user User @relation(...) stays a
relation.
Table names follow the provider too. SQL Server is generated as [dbo].[table], as
before; every other provider gets the bare table, because the brackets are SQL Server
syntax — [dbo].[users] is invalid in PostgreSQL and fails on SQLite with “no such
table: dbo.users”. The adapter quotes the name for whichever dialect it is connected
to. A model that declares @@schema("main") keeps that prefix everywhere.
| Type | Description | Example | TypeScript |
|---|---|---|---|
NVARCHAR(n) |
Variable-length Unicode string | NVARCHAR(255) |
string |
VARCHAR(n) |
Variable-length ASCII string | VARCHAR(100) |
string |
CHAR(n) |
Fixed-length string | CHAR(10) |
string |
NCHAR(n) |
Fixed-length Unicode string | NCHAR(10) |
string |
TEXT |
Large text field | TEXT |
string |
NTEXT |
Large Unicode text field | NTEXT |
string |
XML |
XML data | XML |
string |
INT |
32-bit integer | INT |
number |
BIGINT |
64-bit integer | BIGINT |
number \| bigint |
SMALLINT |
16-bit integer | SMALLINT |
number |
TINYINT |
8-bit integer | TINYINT |
number |
FLOAT |
Floating point | FLOAT |
number |
REAL |
Single-precision float | REAL |
number |
DECIMAL(p,s) |
Fixed precision | DECIMAL(10,2) |
number |
NUMERIC(p,s) |
Fixed precision | NUMERIC(10,2) |
number |
MONEY |
Currency amount | MONEY |
number |
SMALLMONEY |
Small currency amount | SMALLMONEY |
number |
BIT |
Boolean | BIT |
boolean |
DATE |
Date only | DATE |
Date |
DATETIME |
Date and time | DATETIME |
Date |
DATETIME2 |
High precision datetime | DATETIME2 |
Date |
SMALLDATETIME |
Short datetime | SMALLDATETIME |
Date |
DATETIMEOFFSET |
Datetime with timezone | DATETIMEOFFSET |
Date |
TIME |
Time only | TIME |
Date |
UNIQUEIDENTIFIER |
UUID/GUID | UNIQUEIDENTIFIER |
string |
VARBINARY(n) |
Variable binary data | VARBINARY(255) |
Buffer |
BINARY(n) |
Fixed binary data | BINARY(16) |
Buffer |
IMAGE |
Large binary data | IMAGE |
Buffer |
ROWVERSION |
Row version timestamp | ROWVERSION |
Buffer |
VECTOR |
Vector embedding | VECTOR(1536) |
number[] \| string |
GEOGRAPHY |
Spatial data | GEOGRAPHY |
string |
GEOMETRY |
Spatial data | GEOMETRY |
string |
HIERARCHYID |
Hierarchy tree position | HIERARCHYID |
string |
SQL_VARIANT |
Any SQL data type | SQL_VARIANT |
any |
| Type | Description | Example | TypeScript |
|---|---|---|---|
VARCHAR(n) |
Variable-length string | VARCHAR(255) |
string |
CHAR(n) |
Fixed-length string | CHAR(10) |
string |
TEXT |
Large text field | TEXT |
string |
UUID |
Universally unique identifier | UUID |
string |
INTEGER |
32-bit signed integer | INTEGER |
number |
INT4 |
32-bit signed integer (alias) | INT4 |
number |
BIGINT |
64-bit signed integer | BIGINT |
number \| bigint |
INT8 |
64-bit signed integer (alias) | INT8 |
number \| bigint |
SMALLINT |
16-bit signed integer | SMALLINT |
number |
INT2 |
16-bit signed integer (alias) | INT2 |
number |
SERIAL |
Autoincrementing integer | SERIAL |
number |
BIGSERIAL |
Autoincrementing 64-bit integer | BIGSERIAL |
number \| bigint |
SMALLSERIAL |
Autoincrementing 16-bit integer | SMALLSERIAL |
number |
DECIMAL(p,s) |
Exact numeric | DECIMAL(10,2) |
number |
NUMERIC(p,s) |
Exact numeric (alias) | NUMERIC(10,2) |
number |
REAL |
Single precision float | REAL |
number |
DOUBLE PRECISION |
Double precision float | DOUBLE PRECISION |
number |
MONEY |
Currency amount | MONEY |
number |
BOOLEAN |
Logical boolean | BOOLEAN |
boolean |
BOOL |
Logical boolean (alias) | BOOL |
boolean |
DATE |
Calendar date | DATE |
Date |
TIME |
Time of day | TIME |
Date |
TIMETZ |
Time with time zone | TIMETZ |
Date |
TIMESTAMP |
Date and time | TIMESTAMP |
Date |
TIMESTAMPTZ |
Timestamp with time zone | TIMESTAMPTZ |
Date |
INTERVAL |
Time span | INTERVAL |
string |
JSON |
Textual JSON | JSON |
any |
JSONB |
Binary JSON | JSONB |
any |
BYTEA |
Binary data | BYTEA |
Buffer |
INET |
IPv4 or IPv6 host address | INET |
string |
CIDR |
IPv4 or IPv6 network spec | CIDR |
string |
VECTOR |
pgvector embedding | VECTOR(1536) |
number[] \| string |
GEOGRAPHY |
PostGIS geography | GEOGRAPHY |
string |
GEOMETRY |
PostGIS geometry | GEOMETRY |
string |
| Type | Description | Example | TypeScript |
|---|---|---|---|
VARCHAR(n) |
Variable-length string | VARCHAR(255) |
string |
CHAR(n) |
Fixed-length string | CHAR(10) |
string |
NVARCHAR(n) |
National variable-length string | NVARCHAR(255) |
string |
TEXT |
Large text field | TEXT |
string |
TINYTEXT |
Small text field | TINYTEXT |
string |
MEDIUMTEXT |
Medium text field | MEDIUMTEXT |
string |
LONGTEXT |
Very large text field | LONGTEXT |
string |
ENUM(...) |
Enumeration set of values | ENUM('active','inactive') |
string |
SET(...) |
Set of permitted values | SET('read','write') |
string |
INT |
32-bit integer | INT |
number |
INTEGER |
32-bit integer (alias) | INTEGER |
number |
BIGINT |
64-bit integer | BIGINT |
number \| bigint |
SMALLINT |
16-bit integer | SMALLINT |
number |
TINYINT |
8-bit integer | TINYINT |
number |
MEDIUMINT |
24-bit integer | MEDIUMINT |
number |
SERIAL |
Bigint auto-increment alias | SERIAL |
number \| bigint |
FLOAT |
Single-precision float | FLOAT |
number |
DOUBLE |
Double-precision float | DOUBLE |
number |
DECIMAL(p,s) |
Exact fixed-point | DECIMAL(10,2) |
number |
NUMERIC(p,s) |
Exact fixed-point | NUMERIC(10,2) |
number |
BOOLEAN |
Boolean (TINYINT(1)) | BOOLEAN |
boolean |
BOOL |
Boolean alias | BOOL |
boolean |
BIT |
Bit-field flag | BIT |
boolean |
DATE |
Date only | DATE |
Date |
DATETIME |
Date and time | DATETIME |
Date |
TIMESTAMP |
Timestamp with timezone conversion | TIMESTAMP |
Date |
TIME |
Time only | TIME |
Date |
YEAR |
4-digit year | YEAR |
number |
JSON |
Native JSON document | JSON |
any |
BLOB |
Binary large object | BLOB |
Buffer |
TINYBLOB |
Small binary object | TINYBLOB |
Buffer |
MEDIUMBLOB |
Medium binary object | MEDIUMBLOB |
Buffer |
LONGBLOB |
Very large binary object | LONGBLOB |
Buffer |
BINARY(n) |
Fixed-length binary | BINARY(16) |
Buffer |
VARBINARY(n) |
Variable-length binary | VARBINARY(255) |
Buffer |
| Type | Description | Example | TypeScript |
|---|---|---|---|
INTEGER |
Integer value | INTEGER |
number |
INT |
Integer affinity | INT |
number |
BIGINT |
Large integer | BIGINT |
number \| bigint |
TINYINT |
Small integer | TINYINT |
number |
SMALLINT |
Small integer | SMALLINT |
number |
MEDIUMINT |
Medium integer | MEDIUMINT |
number |
REAL |
Floating point value | REAL |
number |
FLOAT |
Floating point affinity | FLOAT |
number |
DOUBLE |
Double precision affinity | DOUBLE |
number |
DECIMAL(p,s) |
Decimal affinity | DECIMAL(10,2) |
number |
NUMERIC |
Numeric affinity | NUMERIC |
number |
BOOLEAN |
Stored as 0 or 1 | BOOLEAN |
boolean |
BOOL |
Boolean alias | BOOL |
boolean |
TEXT |
Text string | TEXT |
string |
VARCHAR(n) |
Text affinity string | VARCHAR(255) |
string |
NVARCHAR(n) |
Text affinity string | NVARCHAR(255) |
string |
CHAR(n) |
Text affinity string | CHAR(10) |
string |
CLOB |
Character large object | CLOB |
string |
DATE |
Stored as ISO text | DATE |
Date |
DATETIME |
Stored as ISO text | DATETIME |
Date |
TIMESTAMP |
Stored as ISO text | TIMESTAMP |
Date |
TIME |
Stored as ISO text | TIME |
Date |
BLOB |
Binary data | BLOB |
Buffer |
UUID |
UUID string | UUID |
string |
JSON |
JSON string | JSON |
any |
VECTOR |
Vector embedding, stored as float32 BLOB | VECTOR(1536) |
number[] \| string |
| Type | Description | Example | TypeScript |
|---|---|---|---|
STRING |
Cell text value | STRING |
string |
VARCHAR |
Text column | VARCHAR |
string |
TEXT |
Long text cell | TEXT |
string |
INT |
Integer cell value | INT |
number |
INTEGER |
Integer cell value | INTEGER |
number |
BIGINT |
Large integer value | BIGINT |
number \| bigint |
FLOAT |
Number cell value | FLOAT |
number |
DOUBLE |
Double number value | DOUBLE |
number |
DECIMAL |
Numeric cell value | DECIMAL |
number |
BOOLEAN |
Checkbox / TRUE/FALSE cell | BOOLEAN |
boolean |
BOOL |
Boolean alias | BOOL |
boolean |
DATE |
Date formatted cell | DATE |
Date |
DATETIME |
Date & time cell | DATETIME |
Date |
BYTES |
Base64 encoded cell | BYTES |
Buffer |
BLOB |
Binary cell data | BLOB |
Buffer |
VECTOR |
Vector embedding string | VECTOR(1536) |
number[] \| string |
Optional Fields
Add ? to make a field optional:
model User {
id NVARCHAR(1000) @id @default(uuid())
name NVARCHAR(255)? // Optional field
email NVARCHAR(255) // Required field
}
Attributes
Primary Key
model User {
id NVARCHAR(1000) @id @default(uuid())
}
Unique Constraint
model User {
email NVARCHAR(255) @unique
}
Default Values
model Post {
id NVARCHAR(1000) @id @default(uuid())
status NVARCHAR(50) @default("draft")
views INT @default(0)
createdAt DATETIME2 @default(now())
}
Description
model User {
id NVARCHAR(1000) @id @default(uuid()) @description("Primary key")
}
Model Directives
Directives are declared at the model level and control table mapping, constraints, and indexes.
Table Mapping
Use @@map() to map a model to a different table name:
model User {
id NVARCHAR(1000) @id @default(uuid())
email NVARCHAR(255)
@@map("app_users")
}
Model Description
model User {
id NVARCHAR(1000) @id @default(uuid())
@@description("User account")
}
Unique Constraints
@@unique() declares a unique constraint across one or more fields. Compound
unique constraints are supported:
model Membership {
id NVARCHAR(1000) @id @default(uuid())
userId NVARCHAR(1000)
orgId NVARCHAR(1000)
@@unique([userId, orgId])
}
Indexes
@@index() declares a database index. Advanced options are supported:
model Order {
id NVARCHAR(1000) @id @default(uuid())
userId NVARCHAR(1000)
total DECIMAL(10,2)
status NVARCHAR(50)
createdAt DATETIME2
// Simple index
@@index([userId])
// Named index (honored in migration diff/generate)
@@index([userId, createdAt], map: "idx_orders_user_created")
// Include columns
@@index([userId], include: [total, status])
// Filtered index
@@index([status], filter: "[status] <> 'cancelled'")
// Index options (e.g. fillfactor)
@@index([userId], options: "fillfactor=80")
}
Supported index options:
| Option | Description | Example |
|---|---|---|
map |
Custom index/constraint name | map: "idx_orders_user" |
include |
Included (non-key) columns | include: [total, status] |
filter |
Filtered index predicate | filter: "[status] <> 'cancelled'" |
options |
Raw index options | options: "fillfactor=80" |
Migrations honor mapped index/unique names, include/filter/options metadata,
and dbo.-qualified table names when comparing schema with the database.
Complete Example
model User {
id NVARCHAR(1000) @id @default(uuid()) @description("Primary key")
email NVARCHAR(255) @unique @description("User email")
name NVARCHAR(255)? @description("Display name")
avatar NVARCHAR(500)? @description("Avatar URL")
role NVARCHAR(50) @default("user") @description("User role")
isActive BIT @default(1) @description("Account status")
createdAt DATETIME2 @default(now()) @description("Creation date")
updatedAt DATETIME2 @default(now()) @description("Last update")
// Relations
posts Post[]
profile Profile?
@@description("User account")
@@map("users")
}
model Post {
id NVARCHAR(1000) @id @default(uuid())
title NVARCHAR(255) @description("Post title")
content TEXT? @description("Post content")
published BIT @default(0) @description("Published status")
authorId NVARCHAR(1000) @description("Author reference")
// Relations
author User @relation(fields: [authorId], references: [id])
tags Tag[]
@@index([authorId])
@@map("posts")
}
model Tag {
id NVARCHAR(1000) @id @default(uuid())
name NVARCHAR(100) @unique
posts Post[]
@@map("tags")
}
Next Steps
- CRUD Operations - Learn how to query your data
- Relations - Define relationships between models
- Advanced Queries - Complex query patterns