Level up your GraphQL + GraphDB knowledge - Register free for Dgraph Day
To optimize your experience, the interactive tutorial is only available on desktop.

Dgraph GraphQL Tour


Hello World


For the following queries to run, you should have completed the introduction to setup your backend, define your schema, and insert the sample data. To learn more, see the GraphQL tour introduction.

Let’s have a look at a “hello world” query in GraphQL.

Each query has at least one named root function, and query results are labeled accordingly.

Dgraph generates these functions based on a schema that you provide. In the GraphQL tour introduction, we looked at an example schema that consisted of two types: Person and Animal. For each type, Dgraph can generate three GraphQL queries, get<Type>, query<Type>, and aggregate<Type>. So, from the two types we provided Dgraph generated the following queries:

  • getPerson
  • queryPerson
  • aggregatePerson
  • getAnimal
  • queryAnimal
  • aggregateAnimal

The getPerson and getAnimal queries require an input identifying the single node of that type to return. If you are new to graph databases, think of GraphQL types as similar to tables, and think of nodes as similar to table rows (or records). The queryPerson and queryAnimal queries return all of the nodes of that type (by default). These queries can be provided inputs to filter, order, and paginate through query results. The aggregatePerson and aggregateAnimal queries return aggregated calculations but do not return the results (nodes) themselves. The most common use cases for aggregate queries is to count the number of nodes. Aggregate queries can accept a filter argument to aggregate a filtered selection of nodes.

Dgraph identifies each node with a unique internal id, its UID. This UID can be mapped to a field in the GraphQL schema by using the ID scalar. In the GraphQL schema, Dgraph also provides a mechanism for referencing nodes by external ids, XID. You can map an XID to a field in the schema by using the @id directive. In our example schema, we mapped the UID to the id field and the XID to the xid field of each type. You do not need to use either of these to perform query<Type> queries, but to use the get<Type> queries you will need to use at least one of these id mechanisms. Because Dgraph maps between UIDs and XIDs, you can use get<Type> queries with either of these identifying fields.

You select fields for your query by providing a list of desired fields in your GraphQL query syntax. In our example query, we request the id, xid, name, and age fields. For any field that does not have a set value, null is returned.

Run the query example provided here to see the results returned.

You will see right away that the results are in JSON format and take the same shape as the query itself.

Something to try: Change the query to select someone else. You can check the sample data again.

The tutorial covers connecting the client to the GraphQL API, so just focus on the query syntax for now.