How do I remove a document with node-mongodb-native driver? - node.js

The operation returns saying deleted: 0
const res = await ctx.db.collection(this.col).removeOne({ _id: ctx.params.id });
Not sure what I'm doing wrong here. GET requests by the { _id: <id> } seem to work fine.
ctx.params.id is defined and is the same as the ObjectId in the database.
According to this doc you can do collection.removeOne() (see example 2) https://mongodb.github.io/node-mongodb-native/2.2/api/Collection.html#remove
// Remove all the document
collection.removeOne({a:1}, {w:1}, function(err, r) {
test.equal(null, err);
test.equal(1, r.result.n);
db.close();
});

Try casting ctx.params.id to ObjectId, which is how mongodb stores identifiers internally.
import { ObjectId } from 'mongodb'
id = ObjectId(ctx.params.id)

Related

Creating unique index via mongodb native driver

There is a code:
const { MongoClient } = require('mongodb')
const db = MongoClient.connect('mongodb://172.17.0.2:27017/test')
db
.then(
async dataBase => {
eduDb = dataBase.db('edu-service-accounts')
const accounts = eduDb.collection('accounts')
await accounts.createIndex({ email: 1 }, { unique: true })
accounts.insertOne({ email: '123' })
}
)
Code above creates an index, but that is no unique. I already read official docs for native mongoDB driver, but can't handle it.
And yes, I've deleted all old indexex before testing that code.
Can someone please show a code that really create an index with unique.
I mean not part of official doc, or something like that - I need code that works.
NOTE: I tested that code with local db and mlab - the same result.
Like the documentation says: db.createIndex(collectionname, index[, options], callback) the creation returns an index. Try to log the result of the callback. Maybe you are getting an error from the db.
Try something like:
// your connection stuff
accounts.createIndex({ email: 1 }, { unique: true }, function(err, result) {
if(err) {
console.log(err);
} else {
console.log(result);
}
});
After that please provide us the logs.

Mongoose findById returns null even with valid id

I have already seen the discussion about the following question with a similar title
mongoose 'findById' returns null with valid id
But my problem is not the database name since all my other connections with the same database in fact the queries on the same collection are working fine.
I am using mongoose 4.13.6, node js 6.11 and mongo 3.4.
It is a post request .
var query=req.body;
I am sending the search parameters as
var findFruit =
{
_id:query._id
}
When I print my findFruit I get :
_id:'5a1cf77920701c1f0aafb85e'
The controller function for this is :
Fruit.findById(findFruit._id,function(err,fruit){
if( _.isNull(err) ){
var response = genRes.generateResponse(true,"found successfully");
callback(response);
}
else{
var response = genRes.generateResponse(false,"there occured some error : "+err);
callback(response);
}
})
I even tried find
Fruit.find(findFruit,function(err,fruit){
if( _.isNull(err) ){
var response = genRes.generateResponse(true,"found successfully");
callback(response);
}
else{
var response = genRes.generateResponse(false,"there occured some error : "+err);
callback(response);
}
})
The collection for sure has the entry under this id .
I went through this git issue as well https://github.com/Automattic/mongoose/issues/3079
Unfortunately I cannot downgrade mongoose as it might affect multiple other working functions.
Edit :
I tried creating ObjectId like :
var mongoose = require('mongoose');
var ObjectID = require('mongodb').ObjectID;
var objectId = new ObjectID();
// Convert the object id to a hex string
var originalHex = objectId.toHexString();
// Create a new ObjectID using the createFromHexString function
var newObjectId = new ObjectID.createFromHexString(query._id);
The model file :
var mongoose = require('mongoose');
var Schema = mongoose.Schema;
var ObjectId = Schema.ObjectId;
var FruitSchema = new Schema({
name : {type : String, unique : true},
color : {type : String}
});
module.exports = mongoose.model('Fruit', FruitSchema);
All my findById("id") calls returned null.
When looking at the collection with Compass I realized that all my _id elements were Strings. Note that the entire collection was imported.
I created a new element in Compass and the _id was created as ObjectId! and when I called findById("id") with that element's id it worked!
My conclusion is that there is obviously a bug with import. I have not found a way to convert the string _id fields to ObjectId's in the actual collection.
All my findById("id") calls returned null, when _id elements are Strings.
In the first place:
Check your mongodb database, if _id is stored as String, findById(id) can not find since it identifies ObjectId. If you've used import database by using mongoimport command and including _id in JSON:
Solution 1:
modify your JSON and for each document, change _id for instance:
_id: "5a68fde3f09ad7646ddec17e" to the following and run mongoimport again:
"_id": { "$oid": "5a68fde3f09ad7646ddec17e" }
Solution 2:
delete _id in the JSON file, drop collection and import again. Mongo will auto-create _id.
After any of solutions above, findById("id") will work.
Secondly:
Specifically in such cases where your _id elements are string, might be a better idea to use mongodb package: npm i mongodb
var MongoClient = require('mongodb').MongoClient;
var url = "mongodb://localhost:27017/";
MongoClient.connect(url, function (err, db) {
if (err) throw err;
var dbo = db.db("your_db_name");
dbo.collection("fruits_collection")
.find({_id:'5a1cf77920701c1f0aafb85e'})
//.find({_id:'5a1cf77920701c1f0aafb85e'}, { projection: { _id: 0, name: 1, color: 1} }) // to select specific columns desired
.toArray(function (err, result) {
if (err) throw err;
console.log(result);
db.close();
});
});
The above simple code, assumed you manage error handling yourself, either through try-catch, or
sending 404 as status code, or redirect to error page template, depending on whether the code is embedded in the Express route handler or not.
Hope this helped .. :)
Still trying to figure out why findById didn't work for me but the following piece of code did it
var mongoose = require('mongoose');
var newObjectId=new mongoose.Types.ObjectId(query._id);
var params={
'_id':newObjectId
}
Fruit.find(params).exec(function (err,fruit) {
if( _.isNull(err) ){
var response = genRes.generateResponse(true,"found successfully");
callback(response);
}
else{
var response = genRes.generateResponse(false,"there occured some error : "+err);
callback(response);
}
})
#Shoom. Yes, worked for me, thanks. findById() expects ObjectID, not a String.
I did not have a constraint to create documents with a specific id, so I imported with no _id. The db had newly-assigned _id as ObjectID.
findById(id), (and updateOne({ _id: id }, ...), started working as expected.

find by _id with Mongoose

I am having trouble with a simple findById with mongoose.
Confirmed the item exists in the DB
db.getCollection('stories').find({_id:'572f16439c0d3ffe0bc084a4'})
With mongoose
Story.findById(topic.storyId, function(err, res) {
logger.info("res", res);
assert.isNotNull(res);
});
won't find it.
I also tried converting to a mongoId, still cannot be found (even though mongoose supposedly does this for you)
var mid = mongoose.Types.ObjectId(storyId);
let story = await Story.findOne({_id: mid}).exec();
I'm actually trying to use this with typescript, hence the await.
I also tried the Story.findById(id) method, still cannot be found.
Is there some gotcha to just finding items by a plain _id field?
does the _id have to be in the Schema? (docs say no)
I can find by other values in the Schema, just _id can't be used...
update: I wrote a short test for this.
describe("StoryConvert", function() {
it("should read a list of topics", async function test() {
let topics = await Topic.find({});
for (let i = 0; i < topics.length; i ++) {
let topic = topics[i];
// topics.forEach( async function(topic) {
let storyId = topic.storyId;
let mid = mongoose.Types.ObjectId(storyId);
let story = await Story.findOne({_id: mid});
// let story = await Story.findById(topic.storyId).exec();
// assert.equal(topic.storyId, story._id);
logger.info("storyId", storyId);
logger.info("mid", mid);
logger.info("story", story);
Story.findOne({_id: storyId}, function(err, res) {
if (err) {
logger.error(err);
} else {
logger.info("no error");
}
logger.info("res1", res);
});
Story.findOne({_id: mid}, function(err, res) {
logger.info("res2", res);
});
Story.findById(mid, function(err, res) {
logger.info("res3", res);
// assert.isNotNull(res);
});
}
});
});
It will return stuff like
Testing storyId 572f16439c0d3ffe0bc084a4
Testing mid 572f16439c0d3ffe0bc084a4
Testing story null
Testing no error
Testing res1 null
Testing res2 null
Testing res3 null
I noticed that topic.storyId is a string
not sure if that would cause any issues mapping to the other table.
I tried also adding some type defs
storyId: {
type: mongoose.Schema.Types.ObjectId,
required: false
}
Because this query finds the doc in the shell:
db.getCollection('stories').find({_id:'572f16439c0d3ffe0bc084a4'})
That means that the type of _id in the document is actually a string, not an ObjectId like Mongoose is expecting.
To find that doc using Mongoose, you'd have to define _id in the schema for Story as:
_id: { type: String }
If your Mongo schema is configured to use Object Id, you query in nodeJS using
models.Foo.findById(id)
where Foo is your model and id is your id.
here's a working example
router.get('/:id', function(req, res, next) {
var id = req.params.id
models.Foo.findById(id)
.lean().exec(function (err, results) {
if (err) return console.error(err)
try {
console.log(results)
} catch (error) {
console.log("errror getting results")
console.log(error)
}
})
})
In Mongo DB your query would be
{_id:ObjectId('5c09fb04ff03a672a26fb23a')}
One solution is to use mongoose.ObjectId()
const Model = require('./model')
const mongoose = require('mongoose')
Model.find({ id: mongoose.ObjectId(userID) })
It works, but it is weird because we are using id instead of _id
This is how we do it now:
const { mongoose } = require("mongoose");
YourModel.find({ _id: mongoose.Types.ObjectId("572f16439c0d3ffe0bc084a4") });
I got into this scenario too. This was how I solved it;
According to the mongoose documentation, you need to tell mongoose to
return the raw js objects, not mongoose documents by passing the lean option and setting it to true. e.g
Adventure.findById(id, 'name', { lean: true }, function (err, doc) {});
in your situation, it would be
Story.findById(topic.storyId, { lean: true }, function(err, res) {
logger.info("res", res);
assert.isNotNull(res);
});
If _id is the default mongodb key, in your model set the type of _id as this:
_id: mongoose.SchemaTypes.ObjectId
Then usind mongoose you can use a normal find:
YourModel.find({"_id": "5f9a86b77676e180c3089c3d"});
models.findById(id)
TRY THIS ONE .
REF LINK : https://www.geeksforgeeks.org/mongoose-findbyid-function/
Try this
Story.findOne({_id:"572b19509dac77951ab91a0b"}, function(err, story){
if (err){
console.log("errr",err);
//return done(err, null);
}else{
console.log(story);
}
});

Automatically remove referencing objects on deletion in MongoDB

Let's suppose I have a schema like this:
var Person = new Schema({
name: String
});
var Assignment = new Schema({
name: String,
person: ObjectID
});
If I delete a person, there can still be orphaned assignments left that reference a person that does not exist, which creates extraneous clutter in the database.
Is there a simple way to ensure that when a person is deleted, all corresponding references to that person will also be deleted?
You can add your own 'remove' Mongoose middleware on the Person schema to remove that person from all other documents that reference it. In your middleware function, this is the Person document that's being removed.
Person.pre('remove', function(next) {
// Remove all the assignment docs that reference the removed person.
this.model('Assignment').remove({ person: this._id }, next);
});
If by "simple" you mean "built-in", then no. MongoDB is not a relational database after all. You need to implement your own cleaning mechanism.
The remove() method is deprecated.
So using 'remove' in your Mongoose middleware is probably not best practice anymore.
Mongoose has created updates to provide hooks for deleteMany() and deleteOne().
You can those instead.
Person.pre('deleteMany', function(next) {
var person = this;
person.model('Assignment').deleteOne({ person: person._id }, next);
});
In case if anyone looking for the pre hook but for deleteOne and deleteMany functions this is a solution that works for me:
const mongoose = require('mongoose');
...
const PersonSchema = new mongoose.Schema({
name: {type: String},
assignments: [{type: mongoose.Schema.Types.ObjectId, ref: 'Assignment'}]
});
mongoose.model('Person', PersonSchema);
....
const AssignmentSchema = new mongoose.Schema({
name: {type: String},
person: {type: mongoose.Schema.Types.ObjectId, ref: 'Person'}
});
mongoose.model('Assignment', AssignmentSchema)
...
PersonSchema.pre('deleteOne', function (next) {
const personId = this.getQuery()["_id"];
mongoose.model("Assignment").deleteMany({'person': personId}, function (err, result) {
if (err) {
console.log(`[error] ${err}`);
next(err);
} else {
console.log('success');
next();
}
});
});
Invoking deleteOne function somewhere in service:
try {
const deleted = await Person.deleteOne({_id: id});
} catch(e) {
console.error(`[error] ${e}`);
throw Error('Error occurred while deleting Person');
}
You can leave the document as is, even when the referenced person document is deleted. Mongodb clears references which point to non-existing documents, this doesn't happen immediately after deleting the referenced document. Instead, when you perform action on the document, e.g., update. Moreover, even if you query the database before the references are cleared, the return is empty, instead of null value.
Second option is to use $unset operator as shown below.
{ $unset: { person: "<person id>"} }
Note the use of person id to represent the value of the reference in the query.
you can use soft delete. Do not delete person from Person Collection instead use isDelete boolean flag to true.
Use $pull. Suppose you have a structure like this.
Stuff Collection:
_id: ObjectId('63dd23c633c17a718c4c5db7')
item: "Item 1"
user: ObjectID('63de669153bc12ecb9081b9e')
User collection:
_id: ObjectId('63de669153bc12ecb9081b9e')
stuff: array[ObjectId('63dd23c633c17a718c4c5db7'), ObjectId('63de3a69715ec134e161b0ea')]
Then after you remove the stuff:
const stuff = Stuff.findById(req.params.id)
const user = User.findById(req.params.id)
await stuff.remove()
// here you can use $pull to update
await user.updateOne({
$pull: {
stuff: stuff.id
}
})
you can simply call the model that needs to be deleted and delete that document like this:
PS: This answer is not specific to the question schema.
const Profiles = require('./profile');
userModal.pre('deleteOne', function (next) {
const userId = this.getQuery()['_id'];
try {
Profiles.deleteOne({ user: userId }, next);
} catch (err) {
next(err);
}
});
// in user delete route
exports.deleteParticularUser = async (req, res, next) => {
try {
await User.deleteOne({
_id: req.params.id,
});
return res.status(200).json('user deleted');
} catch (error) {
console.log(`error`, error);
return next(error);
}
};

NodeJS + MongoDB: Getting data from collection with findOne ()

I have a collection "companies" with several objects. Every object has "_id" parameter. I'm trying to get this parameter from db:
app.get('/companies/:id',function(req,res){
db.collection("companies",function(err,collection){
console.log(req.params.id);
collection.findOne({_id: req.params.id},function(err, doc) {
if (doc){
console.log(doc._id);
} else {
console.log('no data for this company');
}
});
});
});
So, I request companies/4fcfd7f246e1464d05000001 (4fcfd7f246e1464d05000001 is _id-parma of a object I need) and findOne returns nothing, that' why console.log('no data for this company'); executes.
I'm absolutely sure that I have an object with _id="4fcfd7f246e1464d05000001". What I'm doing wrong? Thanks!
However, I've just noticed that id is not a typical string field. That's what mViewer shows:
"_id": {
"$oid": "4fcfd7f246e1464d05000001"
},
Seems to be strange a bit...
You need to construct the ObjectID and not pass it in as a string. Something like this should work:
var BSON = require('mongodb').BSONPure;
var obj_id = BSON.ObjectID.createFromHexString("4fcfd7f246e1464d05000001");
Then, try using that in your find/findOne.
Edit: As pointed out by Ohad in the comments (thanks Ohad!), you can also use:
new require('mongodb').ObjectID(req.params.id)
Instead of createFromHexString as outlined above.
That's because _id field in mongo isn't of string type (as your req.params.id). As suggested in other answers, you should explicitly convert it.
Try mongoskin, you could use it like node-mongodb-native driver, but with some sugar. For example:
// connect easier
var db = require('mongoskin').mongo.db('localhost:27017/testdb?auto_reconnect');
// collections
var companies = db.collection('companies');
// create object IDs
var oid = db.companies.id(req.params.id);
// some nice functions…
companies.findById();
//… and bindings
db.bind('companies', {
top10: function(callback) {
this.find({}, {limit: 10, sort: [['rating', -1]]).toArray(callback);
}
});
db.companies.top10(printTop10);
You can use findById() which will take care of the id conversion for you.
company = Company.findById(req.params.id, function(err, company) {
//////////
});
In case these didn't work for you, this worked for me for accessing a blog post:
const getSinglePost = async (req, res) => {
let id = req.params.id;
var ObjectId = require('mongodb').ObjectId;
const db = await client.db('CMS');
const data = await db.collection("posts").findOne({ _id: ObjectId(id) })
if (data) {
res.status(200).send(data)
} else res.status(400).send({ message: "no post found" })
}

Resources