I need to set a field from another collection as a foreign key to my mongoose schema.
I have a counter pid to productSchema. I need to set pid as a foreignkey to orderSchema
const orderSchema = mongoose.Schema({
id: mongoose.Schema.Types.ObjectId,
product: { type: mongoose.Schema.Types.ObjectId, ref: 'Product',
required:
true },
quantity: { type: Number, default: 1 }
});
const productSchema = mongoose.Schema({
_id: mongoose.Schema.Types.ObjectId,
name: { type: String, required: true },
price: { type: Number, required: true },
productImage: { type: String }
});
Well MongoDB is not a relational DB so it does not support concepts like foreign key. In mongoDB you have to structure your data in such a way that you can fetch information easily.
You should see if you can go with Aggregation or something but mongoDB does not have foreign keys. More important you should make correct choice when you are creating a DB , you should select appropriate DB according to your requirements.
Related
I'm working on an e-commerce project in Express and MongoDB. I'm confused with architecture on how to make relationship between two models such that if I delete one element from a table, all of it's associations should be deleted. CASCADE delete, if I'm using the correct term.
I'm not a database designer, just learning Express and MongoDB. So apologies if my schema is not that good.
I have two tables (or documents in MongoDB). One is Order with schema below:
const orderSchema = new mongoose.Schema({
shippingInfo : {
type: mongoose.Types.ObjectId,
ref: 'Address'
},
user : {
type: mongoose.Types.ObjectId,
ref: 'User',
required: true
},
orderItems: [
{
type: mongoose.Types.ObjectId,
ref:'OrderItem'
}
],
totalPrice: {
type: Number,
required: true,
default: 0.0
},
status: {
type: String,
enum: ['processing', 'shipped', 'delivered','cancelled'],
default: 'processing'
},
deliveredAt: {
type: Date,
}
})
and OrderItems
const orderItemSchema = new mongoose.Schema({
product: {
type: mongoose.Types.ObjectId,
ref: 'Product'
},
name: {
type: String,
required: true
},
quantity: {
type: Number,
required: true
},
price: {
type: Number,
required: true
},
image: {
type: String,
required: true
},
})
I want if I delete an Order, all of its OrderItems should be deleted right away (using remove middleware in Order).
I know that Django has something called on_delete=model.CASCADE when we create relationships, but I'm unaware of such thing in Mongoose.
I don't want to explicitly make another API request to search for and delete all OrderItems that are referenced in orderItems array in an Order, once it is deleted. There has to be a better approach for this.
Another post on Stack Overflow suggested that in remove middleware of Order I should do something like
OrderItem.find({ order_id: this._id }).remove().exec();
That would require me to refer order_id in OrderItem right?
And this would create circular dependency since OrderItem would require Order to be created first and vice versa.
What should I do here? Should I change the schema for both tables i.e. remove orderItems entry from Order and instead add order_id in OrderItem? Or is there a Mongoose way to overcome this situation?
I had a mongoose schema like below:
const orderSchema = new Schema({
orderBy: { type: Schema.Types.ObjectId, ref: 'user', required: true },
items: { type: Schema.Types.ObjectId, ref: 'cart', required: true },
...
)}
I have added some data in the database using this schema. After a few days I added a new field like below:
prepareTime: { Type: Number, default: 0 }
After adding this field, I updated all the existing documents with value prepareTime:0
Now, when I save a new document, the prepareTime field with default value 0 is not being saved. How to achieve this functionality?
I started learning some NodeJS, and how to make a REST API from Academind on YouTube, and learned what a relational and non-relational database is, etc.
With MongoDB, writes are rather cheap, so I want to minimize the number of reads that I do. At the moment I am trying to see how I could make an API, that will be for an app that's similar to discord's, although it'll be for fun.
Is this the right way to make a Schema?
const mongoose = require('mongoose')
const userSchema = mongoose.Schema({
_id: mongoose.Schema.Types.ObjectId,
name: { type: String, required: true, unique: true},
email: { type: String, required: true },
password: { type: String, required: true }, // TODO: Hashing, etc
guilds: [{
_id: mongoose.Schema.Types.ObjectId,
name: {type: String, required: true},
channels: [{
_id: mongoose.Schema.Types.ObjectId,
name: {type: String, required: true},
// Only the X most recent messages
messages: [{
_id: mongoose.Schema.Types.ObjectId,
message: {type: String, required: true},
user: {
_id: mongoose.Schema.Types.ObjectId,
name: {type: String, required: true}
}
}]
}],
// Only an X amount of users
users: [{
_id: mongoose.Schema.Types.ObjectId,
name: {type: String, required: true}
}]
}]
})
module.exports = mongoose.model('User', userSchema)
And then for the Guilds,
const mongoose = require('mongoose')
const guildSchema = mongoose.Schema({
_id: mongoose.Schema.Types.ObjectId,
name: {type: String, required: true},
channels: [{
_id: mongoose.Schema.Types.ObjectId,
name: {type: String, required: true},
// Only an X amount of messages
messages: [{
_id: mongoose.Schema.Types.ObjectId,
message: {type: String, required: true},
user: {
_id: mongoose.Schema.Types.ObjectId,
name: {type: String, required: true}
}
}]
}],
// All the users
users: [{
_id: mongoose.Schema.Types.ObjectId,
name: {type: String, required: true}
}]
})
module.exports = mongoose.model('Guild', guildSchema)
Channel Schema
const mongoose = require('mongoose')
const channelSchema = mongoose.Schema({
_id: mongoose.Schema.Types.ObjectId,
name: { type: String, required: true },
guild: {
_id: mongoose.Schema.Types.ObjectId,
name: { type: String, required: true },
channels: [{
_id: mongoose.Schema.Types.ObjectId,
name: { type: String, required: true }
}],
// The users of the guild, or just the channel?
// Could add a users object outisde of the guild object
users: [{
_id: mongoose.Schema.Types.ObjectId,
name: { type: String, required: true }
}]
}
})
module.exports = mongoose.model('Channel', channelSchema)
And finally for the messages
const mongoose = require('mongoose')
const messageSchema = mongoose.Schema({
_id: mongoose.Schema.Types.ObjectId,
user: {
_id: mongoose.Schema.Types.ObjectId,
name: {type: String, required: true}
},
message: {type: String, required: true},
channel: {
guild: {
_id: mongoose.Schema.Types.ObjectId,
name: {type: String, required: true}
// Store more data for each message?
}
}
})
module.exports = mongoose.model('Message', messageSchema)
I am not sure if this is how a non-relational schema should look like. If it's not, how would I go about to store the data that I need?
Also, let's say that I POST a message on channel X on guild Y with the users A B and C, how would I go about to update all the entries, to add a message?
I've only used the User.find({_id: id}).exec().then().catch() so far, so I am not sure how to go about to update them.
Thanks in advance!
The messages collection should be on its own, do not embed it into any collection. This is not a good idea to embed data that will grow without limit.
The idea to store the last 5 messages into other collection looks painful to implement.
Embed denormalised data from all collections into the users collection seems like a problem when you will have to update guilds, channels, guilds users.
You may embed channels into guilds. Channels would not grow without a limit, should be a reasonable amount, less than 100 of channels per guild and probably it always used with a guild that they belong to. If not, consider not to embed channels into guilds.
The power of mongodb is to build the schema that reflects how your app is using data. I would recommend starting with normalized data. And when problems with creating, reading, updating, deleting data will occur then make appropriate changes in your mongoose schema to solve the problem. Premature optimization will only hurts in the long run.
As always an answer depends on details. Since I do not know all details I would recommend three part article by William Zola, Lead Technical Support Engineer at MongoDB. part 1 part 2 part 3
Is there a way in mongoose + Node.js/Express to define the relation between the foreign key field and what I refer to that field in the model is? My issue is that I have a mongo database where my foreign keys are all formatted like 'exampleId' instead of 'example'. I could just call out 'exampleId' directly but then it leads to weird things like when I populate 'exampleId' instead of 'example' (which is confusing because once populated, it is now the 'example' itself instead of its id).
Here is how I do it now and it works with my graphQL server, but only if my field in the database is 'course' while my database's field is 'courseId'
const CourseSchema = new Schema({
_id: { type: String },
sections: [{
type: Schema.Types.String,
ref: 'Section'
}],
});
const SectionType = new GraphQLObjectType({
name: 'SectionType',
fields: () => ({
id: { type: GraphQLID },
courseId: {
type: require('./course_type'),
resolve(parentValue) {
return Section.findById(parentValue)
.populate('course')
.then(section => section.course);
}
},
}),
});
I figured it out! With the newest version of mongoose, you actually can use virtual fields to accomplish what I wanted to do and this technique allows for flexibility in laying out your schema. Say that my MongoDB collections look like the following:
Courses { _id, sectionIds }
Lectures { _id, courseId }
I can use the following schema in mongoose and it will allow me to refer to course.lectures or lecture.course instead of the usual course.lectureIds or section.courseId:
const CourseSchema = new Schema({
_id: { type: String },
});
CourseSchema.virtual('sections', {
type: Schema.Types.String,
ref: 'Section',
localField: 'sectionIds',
foreignField: '_id',
justOne: false,
});
CourseSchema.statics.findSections = function(id) {
return this.findById(id)
.populate('sections')
.then(course => course.sections);
}
const SectionSchema = new Schema({
_id: { type: String },
});
SectionSchema.virtual('course', {
type: Schema.Types.String,
ref: 'Course',
localField: 'courseId',
foreignField: '_id',
justOne: true,
});
Actually MongoDB isn't a relational database. You can alter the field and its name whatever you like. Ex I Have an Owner(Meteor.users) table and Patient Table with this column
ownerid : {type: String, min: 1},
firstname: {type: String, min: 1},
lastname: {type: String, min: 1},
middlename: {type: String, min: 1, optional: true},
createdbyid: { type: String },
createdbyname: { type: String },
createdat: { type: Date, defaultValue: new Date() },
updatedbyid: { type: String, optional: true },
updatedbyname : { type: String, optional: true },
updatedat: { type: Date, defaultValue: new Date() },
I can easily stamp the value of my {Meteor.Users()._id} to ownerid of my designated patient by just processing them at meteor.methods. You don't have to worry about foreign keys mongo doesn't do relational databases you can customize your database whatever you like. I Hope this helps ;)
Mongoose Documentation posits that _id has to be used in refs and that[i]t is important to match the type of _id to the type of ref. , e.g.:
var personSchema = Schema({
_id : Number, //it is `Number`
name : String,
age : Number,
stories : [{ type: Schema.Types.ObjectId, ref: 'Story' }]
});
var storySchema = Schema({
_creator : { type: Number, ref: 'Person' },
title : String,
fans : [{ type: Number, ref: 'Person' }] // use `Number` to match
});
I also wonder if by 'database' you mean 'collection'.
Suppose I have two schemas:
1. UserEnrolledCourses
var userCoursesSchema = new mongoose.Schema({
user: { type: mongoose.Schema.Types.ObjectId, ref: 'users'},
courseId: { type: mongoose.Schema.Types.ObjectId, ref: 'courses'},
isEnrolled: Boolean,
});
2. CourseResources
var resourcesSchema = new mongoose.Schema({
courseId: { type: mongoose.Schema.Types.ObjectId, ref: 'courses', required: true },
type: {type:String, required:true},
});
Both of them shared the same courseId reference from courses schema.
So, my aim is to generate result from query that for each courseId that one user enrolled, list all of the resources that available. Is that possible?
In mongoDB, you are performing queries on one concrete collection. The only exception is the left outer join with new method $lookup in aggregation for mongodb 3.2 and more. Look at documentation