GraphQL vs Gel

Despite being fundamentally different technologies, Gel and GraphQL solve some of the same problems. GraphQL is a specification for implementing typesafe APIs, consisting of a data modeling language, a query language, and a set of associated tools. Gel is a database, coupled with an associated data modeling language, a query language (EdgeQL), and set of associated tools.

Indeed GraphQL was a source of inspiration for the design of Gel, specifically:

its succinct, declarative, object-oriented modeling language
its composable, syntax-rich query language
its intuitive approach to field selection and deep fetching

So below we compare GraphQL and Gel side-by-side. If you're familiar with GraphQL, this should be an effective way to map syntax and concepts you're already familiar with from GraphQL onto Gel. Moreover, the examples below will highlight scenarios where GraphQL falls short relative to a fully-fledged query language like EdgeQL.

Modeling

Both Gel and GraphQL provide a schema definition language (SDL) for defining your schema. In the case of GraphQL it's your API schema; in the case of Gel it's your application's data model. Let's see how these two SDLs stack up side-by-side.

Scalar types

Object types

Abstract types

Relations

Polymorphic types

Unions

Enums and Arrays

Custom scalars

Copy

# No native support for
# temporal or JSON fields

String
Boolean
Int
Float
ID

Copy

str
bool
int{16|32|64}
float{32|64}
uuid
bigint
decimal
sequence
datetime
duration
cal::local_datetime
cal::local_date
cal::local_time
json

By virtue of being a database instead of an API definition language, Gel's SDL is far broader in scope than GraphQL, to include concepts like constraints, default values, computed properties/links, indexes, stored procedures, link properties, and much more. Check out the Data Modeling showcase for a more complete picture of Gel's data modeling capabilities.

Query language

GraphQL and Gel both feature an associated query language. To facilitate side-by-side comparison, we'll be considering a hypothetical GraphQL API that supports querying and mutation of your entire database.

In fact, this isn't just a hypothetical; Gel instances can process EdgeQL queries! With the flip of a switch, you can query and mutate data in Gel via GraphQL. For details, check out the GraphQL showcase page and documentation.

All queries in the following sections assume the following schema:

Copy

abstract type Person {
  required name: str {
    constraint exclusive;
  };
}

type Villain extending Person {
  nemesis: Hero;
}

type Hero extending Person {
  secret_identity: str;
  number_of_movies: int64;
  multi link villains := .<nemesis[IS Villain];
}

Querying

Simple query

Nested query

Basic filter

Nested filters

Advanced filtering

Pagination

Polymorphic query

Variables

Copy

query getHeroes {
  Hero {
    id
    name
    secret_identity
  }
}

Copy

select Hero {
  id,
  name,
  secret_identity
}

Mutation

Insert

Update

Delete

Insert multiple

Nested insert

Nested connect

Disconnect

Mutation + fetch

Copy

mutation {
  insert_Hero(data: {
    name: "The Falcon",
    secret_identity: "Sam Wilson"
  }) { id }
}

Copy

insert Hero {
  name := 'The Falcon',
  secret_identity := 'Sam Wilson'
};

The examples above draw comparisons between GraphQL and Gel where possible. However, due to the fundamentally different nature of the technologies, there is a wide swath of Gel features that are out of scope for an API specification language like GraphQL: computed properties, subqueries, with clauses, type casting, and a large set of built-in operators and utility functions. For a more comprehensive overview of Gel features and syntax check out the EdgeQL showcase.

Learn Gel

5-min Quickstart

Interactive EdgeQL Tutorial