Terminology​

Below is an overview of EdgeDB’s terminology. Use it as a roadmap, but don’t worry if it doesn’t all make sense immediately. The following pages go into detail on each concept below.

Instance

An EdgeDB instance is a collection of databases that store their data in a shared directory and are managed by a running EdgeDB process. You can create, start, stop, and destroy instances on your local computer with the EdgeDB CLI. Instances listen for incoming queries on a connection port.

Database

Each instance can contain several databases, each with a unique name. At the time of creation, all instances contain a single database called edgedb. This is the default database; all incoming queries are executed against it unless otherwise specified.

Module

Each database can contain several modules, each with a unique name. Modules can be used to organize large schemas into logical units. In SDL, module blocks are used to declare types inside a particular module.

module default {
  # declare types here
}
module another_module {
  # more types here
}

Some module names (std, math, cal, schema, sys, cfg) are reserved by EdgeDB and contain pre-defined types, utility functions, and operators. It’s common to declare an application’s entire schema inside a single module called default.

Object Type

Schemas are predominantly composed of object types, the EdgeDB equivalent of SQL tables. Object types contain properties and links. Both properties and links are associated with a unique name (e.g. first_name, friends, etc) and a cardinality, which can be either single (the default) or multi. Object types can be augmented with constraints, annotations, and indexes.

Property

Properties correspond to either a scalar type (e.g. str, int64) or a collection type (an array or a tuple). They can be augmented with constraints, annotations, and default values. They can also be marked as readonly.

Link

Links represent relationships between object types. Like properties, they can marked as readonly and augmented with constraints, annotations, and default values. They can optionally contain link properties.

Computed

Links and properties can also be computed. Computed links and properties are not physically stored in the database, but they can be used in queries just like non-computed ones. The value will be computed as needed.

Function

You can also declare custom strongly-typed functions in addition to the large library of built-in functions and operators.