Create object types
The most basic components of a GraphQL schema are object types, which just
represent a kind of object you can fetch from your service, and specify the
fields contained in the object.
Object types also have
directives, which
are identifiers preceded by an @ character.
Directives are GraphQL server-specific, and implemented in the server.
Stargate implements directives that are specific to the Apache Cassandra
backend, as well as Apollo data federation-specific directives that allow Stargate
to work with the Apollo gateway.
-
Book type
-
Reader type
type Book @key @cql_entity(name: "book") @cql_input { (1) (2) (3)
title: String! @cql_column(partitionKey: true, name: "book_title") (4) (5)
isbn: String @cql_column(clusteringOrder: ASC) (6)
author: [String] @cql_index(name: "author_idx", target: VALUES) (7)
}type Reader @key @cql_entity(name: "reader") @cql_input {
name: String! @cql_column(partitionKey: true)
user_id: Uuid! @cql_column(clusteringOrder: ASC)
birthdate: Date @cql_index(name: "date_idx") (8)
email: [String] @cql_column(typeHint: "set<varchar>") (9)
reviews: [Review] @cql_index(name: "review_idx", target: VALUES)
address: [Address]
}Directives in example:
| 1 | Book: @key defines the fields that uniquely identify an object.
Stargate assigns primary key fields automatically as @key fields.
For other servers, the fields must be identified in a string, such as @key(fields: "title isbn").
Required for Apollo data federation. |
| 2 | Book: @cql_entity(name: "book") defines an alternate table name for the underlying database.
Cassandra, Astra DB, and DataStax Enterprise all require double-quotes around table names
that are CamelCase.
Changing the name to lowercase makes the object name both GraphQL-friendly (Book) and CQL-friendly (book).
Not required. |
| 3 | Book: @cql_input generates a BookInput type with the same fields as the type Book
in the GraphQL schema, without duplicating the object type in the schema.
Not required but strongly suggested to make writing queries and mutations easier. |
| 4 | Book: @cql_column(partitionKey: true) sets the field to a partition key in the underlying database table.
This directive must be used for each field that comprises the partition key.
IMPORTANT: If no field has a partitionKey: true, but the first field is of data type
ID, Uuid, or TimeUuid, then that field is used as the partition key in the database.
Information about partition keys and clustering keys can be found in the
CQL reference. |
| 5 | Book: @cql_column(partitionKey: true, name: "book_title") also sets the field name to an
alternate value that is CQL-friendly. Note the comma-separated items in @cql_column.
Column renaming is not required. |
| 6 | Book: @cql_column(clusteringOrder: ASC) sets the field to a clustering column in the underlying database table.
This directive must be used for each field that is a clustering column.
The default clustering order is ASC, or ascending.
Not required. |
| 7 | Book: @cql_index(name: "author_idx", target: VALUES) creates an index of the list of author names,
using the values listed in the array.
Note that the parameter target is required for arrays (set, list).
Required if queries will use the field as a parameter to retrieve data. |
| 8 | Reader: @cql_index(name: "date_idx") creates an index of the birthdates.
Note that no target is required.
The @cql_index is discussed in Create indexes.
Not required. |
| 9 | Reader: @cql_column(typeHint: "set<varchar>") sets the column data type to a set.
The only valid defaults are set or making a CQL type frozen.
Required if the field is a set or list. |
The fields author, reviews, address are defined as arrays by the addition of square brackets,
reviews: [Review].
Note that reviews and addresses are arrays of UDTs
whereas author is a list of Strings to name each author.