Feathers database adapter for Objection.js, an ORM based on KnexJS SQL query builder for Postgres, MSSQL, MySQL, MariaDB, SQLite3, and Oracle.
npm install --save feathers-objection
npm install --save objection
npm install --save knex
Then add one of the following:
npm install --save pg
npm install --save sqlite3
npm install --save mysql
npm install --save mysql2
npm install --save oracle
npm install --save mssql
If you want to use a MariaDB instance, you can use the mysql driver.
Use feathers generate service
command to generate a new Objection
service.
Please refer to the Feathers database adapter documentation for more details or directly at:
- Querying - The common adapter querying mechanism
- Pagination and Sorting - How to use pagination and sorting for the database adapter
- Extending - How to extend a database adapter
Refer to the official Objection.js documention.
It works like the Knex service adapter, except it has all the benefits of the Objection ORM.
config/defaults.json
{
"mysql": {
"client": "mysql2",
"connection": {
"host": "mysql.example.com",
"user": "root",
"password": "secret",
"database": "example"
}
}
}
objection.js
const { Model } = require('objection');
module.exports = function(app) {
const { client, connection } = app.get('mysql');
const knex = require('knex')({ client, connection, useNullAsDefault: false });
Model.knex(knex);
app.set('knex', knex);
};
model
(required) - The Objection model definitionid
(optional, default:model.idColumn
or'id'
) - The name of the id field property. Use array of strings for composite primary keysevents
(optional) - A list of custom service events sent by this servicepaginate
(optional) - A pagination object containing adefault
andmax
page sizemulti
(optional) - Allowcreate
with arrays andupdate
andremove
withid
null
to change multiple items. Can betrue
for all methods or an array of allowed methods (e.g.[ 'remove', 'create' ]
)whitelist
(optional) - A list of additional query operators to allow (e.g.[ '$eager', '$joinRelation' ]
)
Starting at version 2.0.0 feathers-objection
converts queries securely. If you
want to support additional Objection operators, the whitelist
service option
can contain an array of additional allowed operators. By default, supported
operators are:
'$eq',
'$ne',
'$gte',
'$gt',
'$lte',
'$lt',
'$in',
'$nin',
'$like',
'$notLike',
'$ilike',
'$notILike',
'$or',
'$and'
Eager queries is one way of solving the SQL database relational model in Feathers services, instead of relying with hooks.
Note that all this eager related options are optional.
allowedEager
- relation expression to limit the allowed eager queries in the service. Defaults to'[]'
, meaning no eager queries allowed. SeeallowEager
documentation.eagerFilters
- option to impose compulsory eager filter. It takes an object or array of objects with the following properties:expression
- the relation expression that the filter will be applied.filter
- the filter function. It usesfilterEager
internally.
namedEagerFilters
- object containing named eager filter functions. Filter is opt-in via$eager
parameter.
-
$eager
- eager load relations defined in models'relationMappings
getter methods or in thenamedEagerFilters
option. Seeeager
documentation. -
$joinRelation
- filter based on a relation's field. SeejoinRelation
documentation. -
$joinEager
- filter based on a relation's field usingJoinEagerAlgorithm
. See$joinEager
documentation. -
$pick
- pick properties from result models. Seepick
documentation. -
$select
- add SELECT statement with given array of column names. See$select
documentation. -
$noSelect
- skips SELECT queries in create, patch & remove requests. response data will be based on the input data. -
$between
- filter based on if a column value is between range of values -
$notBetween
- filter based on if a column value is not between range of values -
$like
- filter column value based on a LIKE pattern -
$notLike
- filter column value based on a NOT LIKE pattern -
$regexp
- filter column value based on a REGEXP pattern -
$notRegexp
- filter column value based on a NOT REGEXP pattern -
$ilike
- (Postgres) filter column value based on a case-insensitive LIKE pattern -
$notILike
- (Postgres) filter column value based on a case-insensitive NOT LIKE pattern -
$iRegexp
- (Postgres) filter column value based on a case-insensitive REGEXP pattern -
$notIRegexp
- (Postgres) filter column value based on a case-insensitive NOT REGEXP pattern -
$containsKey
(Postgres) - filter based on if a column contains a key -
$any
(Postgres) - filter based on if a column contains any key from array of strings -
$all
(Postgres) - filter based on if a column contains all keys from array of strings -
$contains
(Postgres) - filter based on if a column contains a value -
$contained
(Postgres) - filter based on if a column is contained in a value
transaction
- A transaction object. Seetransaction
documentation.mergeAllowEager
- Just like allowEager but instead of replacing query builder’s allowEager expression this method merges the given expression to the existing expression. SeemergeAllowEager
documentation.
Composite primary keys can be passed as the id
argument using the following
methods:
- String with values separated by the
idSeparator
property (order matter, recommended for REST) - JSON array (order matter, recommended for internal service calls)
- JSON object (more readable, recommended for internal service calls)
When calling a service method with the id
argument, all primary keys are
required to be passed.
idSeparator
- (optional) separator char to separate composite primary keys in theid
argument of get/patch/update/remove external service calls. Defaults to','
.
app.use('/user-todos', service({
id: ['userId', 'todoId'],
idSeparator: ','
})
app.service('/user-todos').get('1,2')
app.service('/user-todos').get([1, 2])
app.service('/user-todos').get({ userId: 1, todoId: 2 })
JSON column will be automatically converted from and to JS object/array and will be saved as text in unsupported databases.
Query against a JSON column in PostgresSQL:
app.service('companies').find({ query: { obj: { numberField: 1.5 } } });
app
.service('companies')
.find({ query: { obj: { numberField: { $gt: 1.5 } } } });
app.service('companies').find({
query: { obj: { 'objectField.object': 'string in obj.objectField.object' } }
});
app.service('companies').find({
query: {
obj: { 'arrayField(0).object': 'string in obj.arrayField[0].object' }
}
});
app.service('companies').find({
query: {
arr: { '(0).objectField.object': 'string in arr[0].objectField.object' }
}
});
Arbitrary relation graphs can be upserted (insert + update + delete) using the
upsertGraph method. See
examples
for a better
explanation.
Runs on update
and patch
service methods when id
is set.
The relation being upserted must also be present in allowedEager
option and
included in $eager
query when using the update
service method.
allowedUpsert
- relation expression to allow relations to be upserted along with update. Defaults tonull
, meaning relations will not be automatically upserted unless specified here. SeeallowUpsert
documentation.upsertGraphOptions
- SeeupsertGraphOptions
documentation.createUseUpsertGraph
- If set totrue
, Graph Upsert will also be used for.create(data, params)
method instead of Graph Insert.
app.use('/companies', service({
model: Company,
allowedEager: 'clients',
allowedUpsert: 'clients'
})
app.service('/companies').update(1, {
name: 'New Name',
clients: [{
id: 100,
name: 'Existing Client'
}, {
name: 'New Client'
}]
})
In the example above, we are updating the name of an existing company, along with adding a new client which is a relationship for companies. The client without the ID would be inserted and related. The client with the ID will just be updated (if there are any changes at all).
Arbitrary relation graphs can be inserted using the insertGraph method. Provides
the ability to relate the inserted object with its associations. Runs on the
.create(data, params)
service method.
The relation being created must also be present in allowedEager
option and
included in $eager
query.
allowedInsert
- relation expression to allow relations to be created along with insert. Defaults tonull
, meaning relations will not be automatically created unless specified here. SeeallowInsert
documentation.insertGraphOptions
- SeeinsertGraphOptions
documentation.
users.service.js
const createService = require('feathers-objection');
const createModal = require('../../models/users.model');
const hooks = require('./users.hooks');
module.exports = function(app) {
const Modal = createModal(app);
const paginate = app.get('paginate');
const options = {
model: Modal,
paginate,
whitelist: ['$eager', '$joinRelation'],
allowedEager: 'todos'
};
app.use('/users', createService(options));
const service = app.service('users');
service.hooks(hooks);
};
todos.service.js
const createService = require('feathers-objection');
const createModal = require('../../models/todos.model');
const hooks = require('./todos.hooks');
module.exports = function(app) {
const Modal = createModal(app);
const paginate = app.get('paginate');
const options = {
model: Modal,
paginate,
whitelist: ['$eager', '$joinRelation'],
allowedEager: '[user, subtask]',
namedEagerFilters: {
unDone: function(builder) {
builder.where('done', false);
}
},
eagerFilters: [
{
expression: 'subtask',
filter: function(builder) {
builder.where('archived', true);
}
}
]
};
app.use('/todos', createService(options));
const service = app.service('todos');
service.hooks(hooks);
};
Use eager queries as follows:
// Get all todos and their unfinished tasks
app.service('/todos').find({
query: {
$eager: 'subtask(unDone)'
}
});
// Get all todos of an active user with firstName 'John'
app.service('/todos').find({
query: {
'user.firstName': 'John',
$eager: 'user(active)',
$joinRelation: 'user(active)'
}
});
See this article for more information.
Objection requires you to define Models for your tables:
users.model.js
const { Model } = require('objection');
class User extends Model {
static get tableName() {
return 'user';
}
static get jsonSchema() {
return {
type: 'object',
required: ['firstName', 'lastName'],
properties: {
id: { type: 'integer' },
firstName: { type: 'string', maxLength: 45 },
lastName: { type: 'string', maxLength: 45 },
status: {
type: 'string',
enum: ['active', 'disabled'],
default: 'active'
},
address: {
type: 'object',
properties: {
street: { type: 'string' },
city: { type: 'string' },
zipCode: { type: 'string' }
}
},
list: {
type: 'array',
maxItems: 3,
items: { type: 'string' }
}
}
};
}
static get relationMappings() {
const Todo = require('./todos.model')();
return {
todos: {
relation: Model.HasManyRelation,
modelClass: Todo,
join: {
from: 'user.id',
to: 'todo.userId'
}
}
};
}
static get namedFilters() {
return {
active: builder => {
builder.where('status', 'active');
}
};
}
$beforeInsert() {
this.createdAt = this.updatedAt = new Date().toISOString();
}
$beforeUpdate() {
this.updatedAt = new Date().toISOString();
}
}
module.exports = function(app) {
if (app) {
const db = app.get('knex');
db.schema
.hasTable('user')
.then(exists => {
if (!exists) {
db.schema
.createTable('user', table => {
table.increments('id');
table.string('firstName', 45);
table.string('lastName', 45);
table.enum('status', ['active', 'disabled']).defaultTo('active');
table.timestamp('createdAt');
table.timestamp('updatedAt');
})
.then(() => console.log('Created user table'))
.catch(e => console.error('Error creating user table', e));
}
})
.catch(e => console.error('Error creating user table', e));
}
return User;
};
module.exports = User;
todos.model.js
const { Model } = require('objection');
class Todo extends Model {
static setup(app) {
this.app = app;
}
static get tableName() {
return 'todo';
}
static get jsonSchema() {
return {
type: 'object',
required: ['userId', 'text'],
properties: {
id: { type: 'integer' },
userId: { type: 'integer' },
text: { type: 'string', maxLength: 500 },
complete: { type: 'boolean', default: false },
dueDate: { type: 'string', format: 'date-time' }
}
};
}
static get relationMappings() {
const User = require('./users.model')();
return {
user: {
relation: Model.BelongsToOneRelation,
modelClass: User,
join: {
from: 'todo.userId',
to: 'user.id'
}
}
};
}
static get namedFilters() {
const knex = this.app.get('knex');
return {
overdue: builder => {
builder
.where('complete', '=', false)
.where('dueDate', '<', knex.fn.now());
}
};
}
$beforeInsert() {
this.createdAt = this.updatedAt = new Date().toISOString();
}
$beforeUpdate() {
this.updatedAt = new Date().toISOString();
}
}
module.exports = function(app) {
if (app) {
Todo.setup(app);
const db = app.get('knex');
db.schema
.hasTable('todo')
.then(exists => {
if (!exists) {
db.schema
.createTable('todo', table => {
table.increments('id');
table.integer('userId');
table.string('text', 500);
table.boolean('complete');
table.timestamp('dueDate');
table.timestamp('createdAt');
table.timestamp('updatedAt');
})
.then(() => console.log('Created todo table'))
.catch(e => console.error('Error creating todo table', e));
}
})
.catch(e => console.error('Error creating todo table', e));
}
return Todo;
};
Here's a complete example of a Feathers server with a todos
SQLite service:
$ npm install @feathersjs/feathers @feathersjs/express body-parser feathers-objection objection knex sqlite3
app.js
const feathers = require('@feathersjs/feathers');
const express = require('@feathersjs/express');
const rest = require('@feathersjs/express/rest');
const errorHandler = require('@feathersjs/express/errors');
const bodyParser = require('body-parser');
const createService = require('feathers-objection');
const { Model } = require('objection');
const knex = require('knex')({
client: 'sqlite3',
connection: {
filename: './db.sqlite'
},
useNullAsDefault: false
});
// Bind Objection.js
Model.knex(knex);
// Clean up our data. This is optional and is here
// because of our integration tests
knex.schema.dropTableIfExists('todo').then(function() {
console.log('Dropped todo table');
// Initialize your table
return knex.schema.createTable('todo', function(table) {
console.log('Creating todo table');
table.increments('id');
table.string('text');
table.boolean('complete');
table.timestamp('createdAt');
table.timestamp('updatedAt');
});
});
// Create a feathers instance.
const app = express(feathers())
// Enable REST services
.configure(rest())
// Turn on JSON parser for REST services
.use(bodyParser.json())
// Turn on URL-encoded parser for REST services
.use(bodyParser.urlencoded({ extended: true }));
// Create an Objection Model
class Todo extends Model {
static get tableName() {
return 'todo';
}
static get jsonSchema() {
return {
type: 'object',
required: ['text'],
properties: {
id: { type: 'integer' },
text: { type: 'string' },
complete: { type: 'boolean', default: false }
}
};
}
$beforeInsert() {
this.createdAt = this.updatedAt = new Date().toISOString();
}
$beforeUpdate() {
this.updatedAt = new Date().toISOString();
}
}
// Create Objection Feathers service with a default page size of 2 items
// and a maximum size of 4
app.use(
'/todos',
createService({
model: Todo,
id: 'id',
paginate: {
default: 2,
max: 4
}
})
);
// Handle Errors
app.use(errorHandler());
// Start the server
module.exports = app.listen(3030);
console.log('Feathers Todo Objection service running on 127.0.0.1:3030');
Run the example with node app
and go to
localhost:3030/todos.
You should see an empty array. That's because you don't have any Todos yet, but you now have full CRUD for your new todos service!
feathers-objection
2.0.0 comes with important security and usability updates.
Important: For general migration information to the new database adapter functionality see crow.docs.feathersjs.com/migrating.html#database-adapters.
The following breaking changes have been introduced:
- All methods allow additional query parameters
- Multiple updates are disabled by default (see the
multi
option) - Objection related operators are disabled by default (see the
whitelist
option)
Copyright © 2019
Licensed under the MIT license.