This is a simple lib to load data to database using sequelize.
It is intended for easily setting up test data.
Yaml and json formats are supported. Includes a grunt task.
Duplicate records are not inserted.
API returns bluebird promises, but callbacks can also be used as the last argument.
Tested with latest Sequelize (3.24), should work on 2.x.
npm install sequelize-fixtures
npm test
const sequelize_fixtures = require('sequelize-fixtures');
//a map of [model name] : model
//see offical example on how to load models
//https://github.com/sequelize/express-example/blob/master/models/index.js
const models = require('./models');
//from file
sequelize_fixtures.loadFile('fixtures/test_data.json', models).then(function(){
doStuffAfterLoad();
});
//can use glob syntax to select multiple files
sequelize_fixtures.loadFile('fixtures/*.json', models).then(function(){
doStuffAfterLoad();
});
//array of files
sequelize_fixtures.loadFiles(['fixtures/users.json', 'fixtures/data*.json'], models).then(function(){
doStuffAfterLoad();
});
//specify file encoding (default utf8)
sequelize_fixtures.loadFile('fixtures/*.json', models, { encoding: 'windows-1257'}).then(function(){
doStuffAfterLoad();
});
//load fixtures inside a transaction
sequelize.transaction(function(tx) {
sequelize_fixtures.loadFile('fixtures/*.json', models, { transaction: tx}).then(doStuffAfterLoad);
});
//apply transform for each model being loaded
sequelize_fixtures.loadFile('fixtures/*.json', models, {
transformFixtureDataFn: function (data) {
if(data.createdAt
&& data.createdAt < 0) {
data.createdAt = new Date((new Date()).getTime() + parseFloat(data.createdAt) * 1000 * 60);
}
return data;
}
}).then(function() {
doStuffAfterLoad();
});
//modify each model being loaded
sequelize_fixtures.loadFile('fixtures/*.json', models, {
modifyFixtureDataFn: function (data) {
if(!data.createdAt) {
data.createdAt = new Date();
}
return data;
}
}).then(function() {
doStuffAfterLoad();
});
//from array
var fixtures = [
{
model: 'Foo',
data: {
propA: 'bar',
propB: 1
}
},
{
model: 'Foo',
data: {
propA: 'baz',
propB: 3
}
}
];
sequelize_fixtures.loadFixtures(fixtures, models).then(function(){
doStuffAfterLoad();
});
[
{
"model": "Foo",
"data": {
"propA": "bar",
"propB": 1
}
},
{
"model": "Foo",
"data": {
"propA": "baz",
"propB": 3
}
}
]
fixtures:
- model: Foo
data:
propA: bar
propB: 1
- model: Foo
data:
propA: baz
propB: 3
module.exports = [
{
"model": "Foo",
"data": {
"propA": "bar",
"propB": 1
}
},
{
"model": "Foo",
"data": {
"propA": "baz",
"propB": 3
}
}
];
You can specify associations by providing related object id or a where clause to select associated object with. Make sure associated objects are described before associations!
Assuming
Car.belongsTo(Owner);
Owner.hasMany(Car);
Associated entity can be mapped by providing either an id:
[
{
"model": "Owner",
"data": {
"id": 11,
"name": "John Doe",
"city": "Vilnius"
}
},
{
"model": "Car",
"data": {
"id": 203,
"make": "Ford",
"owner": 11
}
}
]
Or a property-value map (like a 'where' object) to find the entity with:
[
{
"model": "Owner",
"data": {
"name": "John Doe",
"city": "Vilnius"
}
},
{
"model": "Car",
"data": {
"make": "Ford",
"owner": {
"name": "John Doe"
}
}
}
]
Assuming
Project.belongsToMany(Person, {through: 'peopleprojects'});
Person.belongsToMany(Project, {through: 'peopleprojects'});
Associated entities can be mapped using the association 'as' name ('people' for Project and 'projects' for People) or 'through' table name, and providing an array of ids:
[
{
"model":"Person",
"data":{
"id":122,
"name": "Jack",
"role": "Developer"
}
},
{
"model":"Person",
"data":{
"id": 123,
"name": "John",
"role": "Analyst"
}
},
{
"model":"Project",
"data": {
"id": 20,
"name": "The Great Project",
"people": [122, 123]
}
}
]
Or an array of property-value mappings to find the entities with:
[
{
"model":"Person",
"data":{
"name": "Jack",
"role": "Developer"
}
},
{
"model":"Person",
"data":{
"name": "John",
"role": "Analyst"
}
},
{
"model":"Project",
"data": {
"name": "The Great Project",
"people": [
{
"name": "Jack"
},
{
"name": "John"
}
]
}
}
]
If you have a custom through model for a m2m relationship, you can set additional attributes using _through
:
ActorsMovies = sequelize.define("ActorsMovies", {
character: {type: DataTypes.STRING}
});
Movie.belongsToMany(Actor, {through: ActorsMovies});
Actor.belongsToMany(Movie, {through: ActorsMovies});
[
{
"model": "Movie",
"data": {
"name": "Terminator"
}
},
{
"model": "Actor",
"data": {
"name": "Arnie",
"movies": [
{
"name": "Terminator",
"_through": {
"character": "T-80"
}
}
]
}
}
]
For each model you can provide build options that are passed to Model.build() and save options that are passed to instance.save(), example:
{
"model": "Article",
"buildOptions": {
"raw": true,
"isNewRecord": true
},
"saveOptions": {
"fields": ["title", "body"]
},
"data": {
"title": "Any title",
"slug": "My Invalid Slug"
}
}
In case you want to detect duplicates based on specific field or fields rather than all fields (for example, don't include entities with the same id, even if other fields don't match), you can speficy these fields with a 'keys' property.
{
"model": "Person",
"keys": ["email"],
"data": {
"name": "John",
"email": "example@example.com"
}
},
{
"model": "Person",
"keys": ["email"],
"data": {
"name": "Patrick",
"email": "example@example.com"
}
}
In this example only John will be loaded
Gruntfile.js:
grunt.initConfig({
fixtures: {
import_test_data: {
src: ['fixtures/data1.json', 'fixtures/models*.json'],
models: function () { //returns mapping model name: model
return require('../models')
},
options: { //specify encoding, optiona. default utf-8
encoding: 'windows-1257'
}
}
}
});
grunt.loadNpmTasks('sequelize-fixtures');