I have the following mongo aggregate query:
return db.collection('projects').aggregate([
{
$match: {
agents: ObjectId(agent)
}
},
{
$lookup: {
from: "agents",
localField: "agents",
foreignField: "_id",
as: "agents"
}
},
{
$lookup: {
from: "roles",
localField: "roles",
foreignField: "_id",
as: "roles"
}
},
{
$lookup: {
from: "agencies",
localField: "agency",
foreignField: "_id",
as: "agency"
}
},
{
$lookup: {
from: "resources",
localField: "roles.applicants",
foreignField: "_id",
as: "roles.applicants"
}
}
])
It works as it should, embedding the proper documents. However, the "password_hash" field is showing for each applicant. I want to remove that field. If I try to project and set roles.applicants.password_hash: 0 I actually end up getting the entire applicant without the password hash, but the rest of the roles fields are no longer there. So I get something that looks like:
roles: {
applicants: {
name: "Josh"
}
}
That should be
roles: {
title: "Super Hero",
applicants: {
name: "Josh"
}
}
I figured it out. Here is how I did it. First, the projection wasn't the issue with why the roles document was missing fields. It was the lookup.
What I did was changed the roles lookup to use the pipeline method on the nested documents and then did another pipeline on applicants, which first matched the applicants from the resource collection and then handled the projection. It looks like this.
{
$lookup: {
from: "roles",
let: { "roles": "$roles" },
pipeline: [
{ $match: { $expr: { $in: [ "$_id", "$$roles" ] } } },
{
$lookup: {
from: "resources",
let: { "applicants": "$applicants" },
pipeline: [
{ $match: { $expr: { $in: [ "$_id", "$$applicants" ] } } },
{
$project: {
first_name: 1
}
}
],
as: "applicants"
}
}
],
as: "roles"
}
}
The entire aggregate looks like this
return db.collection('projects').aggregate([
{
$match: {
agents: ObjectId(agent)
}
},
{
$lookup: {
from: "agents",
localField: "agents",
foreignField: "_id",
as: "agents"
}
},
{
$lookup: {
from: "agencies",
localField: "agency",
foreignField: "_id",
as: "agency"
}
},
{
$lookup: {
from: "roles",
let: { "roles": "$roles" },
pipeline: [
{ $match: { $expr: { $in: [ "$_id", "$$roles" ] } } },
{
$lookup: {
from: "resources",
let: { "applicants": "$applicants" },
pipeline: [
{ $match: { $expr: { $in: [ "$_id", "$$applicants" ] } } },
{
$project: {
first_name: 1
}
}
],
as: "applicants"
}
}
],
as: "roles"
}
}
])
Related
I have two collections. articles and bookmarks.
articles
{
_id: "5faa889ade5e0a6326a873d3",
name: "article 1"
},
{
_id: "5faa889ade5e0a6326a873d",
name: "article 2"
}
bookmarks
{
_id: "5faa889ade5e0a6326a873d1",
user_id: "5fc7b50da483a66a86aa7e9e",
model_id: "5faa889ade5e0a6326a873d3"
}
I want to join article with bookmark. if user bookmarked a article.
what i have tried
const aggregate = await Articles.aggregate([{
$lookup: {
from: "categories",
localField: "category_id",
foreignField: "_id",
as: "category_id"
}
},
{
$lookup: {
from: "bookmarks",
localField: "_id",
foreignField: "model_id",
as: "bookmarks"
}
}
]);
but it will gives all bookmark for the article not only logged in user bookmark. so how can I add a condition.
{ "user_id": objectId(req.user._id) } // logged user id
You can use $lookup with pipeline starting from MongoDB v3.6,
let to pass localField _id as model_id variable, you can use the field inside lookup pipeline using $$ reference,
pipeline to put $match stage and match your required conditions and user_id condition
{
$lookup: {
from: "bookmarks",
let: { model_id: "$_id" },
pipeline: [
{
$match: {
$expr: { $eq: ["$$model_id", "$model_id"] },
user_id: objectId(req.user._id)
}
}
],
as: "bookmarks"
}
}
Other option for MongoDB v3.4,
$filter to iterate loop of bookmarks and get filtered bookmarks on the base of condition
{
$lookup: {
from: "bookmarks",
localField: "_id",
foreignField: "model_id",
as: "bookmarks"
}
},
{
$addFields: {
bookmarks: {
$filter: {
input: "$bookmarks",
cond: { $eq: ["$$this.user_id", objectId(req.user._id)] }
}
}
}
}
You can have nested pipeline inside $lookup,
db.articles.aggregate([
{
$lookup: {
from: "bookmarks",
let: {
article_id: "$_id"
},
pipeline: [
{
$match: {
$expr: {
$and: [
{
$eq: [
"$model_id",
"$$article_id"
]
},
{
$eq: [
"$user_id",
"5fc7b50da483a66a86aa7e9a"
]
}
]
}
}
}
],
as: "bookmarks"
}
}
])
Here's a working playground
I'm trying to set up a query on the mongo that searches in 3 different collections.
Document of client:
{
"_id": ObjectId("1a")
"razaosocial:"32423424",
"prepository": [
{
"$ref": "prepository",
"$id": ObjectId("2a")
}
]
}
Document of prepository:
{
"_id": ObjectId("2a")
"name:"Jonh",
"prepository": {
"$ref": "representative",
"$id": ObjectId("3a")
}
}
Document of representative:
{
"_id": ObjectId("3a")
"name:"Josh"
}
I'm doing it this way, but it doesn't return anything:
db.clients.aggregate(
[
{
$lookup: {
from: 'prepository',
localField: 'prepository',
foreignField: 'id',
as: 'prepository'
}
},
{ $unwind: "$prepository" },
{ $lookup: {
from: 'representative',
localField: 'id',
foreignField: 'prepository._id',
as: 'prepository.repre'
}
},
{ $group: {
_id: "$_id",
client: { $first: "$razaosocial" },
users: {
$push: "$prepository"
}
} }
])
I'm trying to return
{
"_id": "1a"
"razaosocial:"32423424",
"prepository": [
{
"_id": "2a"
"name:"Jonh",
"representative": {
"_id": "3a"
"name:"Josh"
}
}
]
}
I am grateful for any help
You can use nested lookup,
$lookup with prepository collection
$match with prepository's id
$lookup with representative collection
$unwind deconstruct representative array
db.client.aggregate([
{
$lookup: {
from: "prepository",
as: "prepository",
let: { prepository: "$prepository" },
pipeline: [
{ $match: { $expr: { $in: ["$_id", "$$prepository"] } } },
{
$lookup: {
from: "representative",
localField: "representative",
foreignField: "_id",
as: "representative"
}
},
{ $unwind: "$representative" }
]
}
}
])
Playground
Lookup with three collections re-manage data. I got confuse which to use. Please guide on this
db.post.aggregate([
{
$lookup: {
from: "users",
localField: "created_by",
foreignField: "_id",
as: "created_users"
}
},
{
$lookup: {
from: "comments",
let: {
p_id: "$_id"
},
pipeline: [
{
$match: {
$expr: {
$eq: [
"$post_id",
"$$p_id"
]
}
}
}
],
as: "comments"
}
},
{
$lookup: {
from: "users",
localField: "comments.sender_id",
foreignField: "_id",
as: "commented_user"
}
}
])
There are three collection posts, user, and comments we would like to get who has comment for a post, Commented user should come under comments like this
[
{
"_id": ObjectId("5eeb02881982961ada625c7d"),
"commented_user": [
{
"_id": ObjectId("5e4d0973babf2b74ca868f4d"),
"first_name": "James",
"last_name": "Smith",
"timestamp": 1.582106995137e+12
},
{
"_id": ObjectId("5e4d0973babf2b74ca868f6d"),
"first_name": "Alex",
"last_name": "Jimy",
"timestamp": 1.582106995139e+12
}
],
"comments": [
{
"_id": ObjectId("5eeb08e26fb7f270e4077617"),
"date": 1.592461538924e+12,
"post_id": ObjectId("5eeb02881982961ada625c7d"),
"sender_id": ObjectId("5e4d0973babf2b74ca868f4d"),
"text": "Nice ",
"commented_user": {
"_id": ObjectId("5e4d0973babf2b74ca868f4d"),
"first_name": "James",
"last_name": "Smith",
"timestamp": 1.582106995137e+12
},
},
{
"_id": ObjectId("5eeb08e26fb7f270e4077618"),
"date": 1.592461538923e+12,
"post_id": ObjectId("5eeb02881982961ada625c7d"),
"sender_id": ObjectId("5e4d0973babf2b74ca868f6d"),
"text": "Nice One",
"commented_user": {
"_id": ObjectId("5e4d0973babf2b74ca868f6d"),
"first_name": "Alex",
"last_name": "Jimy",
"timestamp": 1.582106995137e+12
},
}
],
"created_by": ObjectId("5e4e74eb380054797d9db623"),
"created_users": [],
"date": 1.589441206774e+12,
"title": "Covid19"
}
]
here is my attempt https://mongoplayground.net/p/UwRjj-er0K5
Please help on this thanks
You can do something like this:
The strategy is $unwinding the results so we can match per comment and then reconstruct the former structure.
db.post.aggregate([
{
$lookup: {
from: "users",
localField: "created_by",
foreignField: "_id",
as: "created_users"
}
},
{
$lookup: {
from: "comments",
let: {
p_id: "$_id"
},
pipeline: [
{
$match: {
$expr: {
$eq: [
"$post_id",
"$$p_id"
]
}
}
}
],
as: "comments"
}
},
{
$unwind: "$comments"
},
{
$lookup: {
from: "users",
localField: "comments.sender_id",
foreignField: "_id",
as: "commented_user"
}
},
{
$unwind: "$commented_user"
},
{
$addFields: {
comments: {
$mergeObjects: [
"$comments",
{
commented_user: "$commented_user"
}
]
}
}
},
{
$group: {
_id: "$_id",
comments: {
$push: "$comments"
},
created_by: {
$first: "$created_by"
},
created_users: {
$first: "$created_users"
},
date: {
$first: "$date"
},
title: {
$first: "$title"
},
}
}
])
Mongo Playground
Starting with your the code you have attempted, you can just move your second $lookup into the first $lookup's pipeline, then $unwind it
db.post.aggregate([
{
$lookup: {
from: "users",
localField: "created_by",
foreignField: "_id",
as: "created_users"
}
},
{
$lookup: {
from: "comments",
let: { p_id: "$_id" },
pipeline: [
{
$match: {
$expr: { $eq: ["$post_id", "$$p_id"] }
}
},
{ // second level $lookup
$lookup: {
from: "users",
localField: "sender_id",
foreignField: "_id",
as: "commented_user"
}
},
{ // $unwind to get single object instead of array
$unwind: "$commented_user"
}
],
as: "comments"
}
},
])
Mongo Playground
Note for OP: There is discussion to get the latest comment instead. With this approach you can $sort and $limit in the outermost $lookup
db.post.aggregate([
{
$lookup: {
from: "users",
localField: "created_by",
foreignField: "_id",
as: "created_users"
}
},
{
$lookup: {
from: "comments",
let: { p_id: "$_id" },
pipeline: [
{
$match: {
$expr: { $eq: ["$post_id", "$$p_id"] }
}
},
{
$sort: { date: -1 } // or { _id: -1 }
},
{
$limit: 1 // limit to only one element
},
{
$lookup: {
from: "users",
localField: "sender_id",
foreignField: "_id",
as: "commented_user"
}
},
{
$unwind: "$commented_user"
}
],
as: "comments"
}
},
{ // // $unwind to get single object instead of array
$unwind: "$comments"
}
])
Mongo Playground
I have an aggregate call as listed below. I believe the massive delay is caused when the roles lookup happens because there are applicants, hired, and rejected people to lookup from the resources table. There are over 10,000 entries in the resource collection so the query is taking about 6 seconds. Is there something I am doing that is incredibly wrong here? I don't see how I can use indexes because all of the lookups are done by with the _id which is already indexed by default.
The applications, hired, and rejected fields are just arrays of object Ids e.g.
applicants: [
ObjectId('asldkajsdlkj'),
ObjectId('asldkjaoksdjak')
]
Any help would be greatly appreciated.It was taking 6 seconds on an M0 instance and is no faster on an M10 instance.
return db.collection('projects').aggregate([
{
$match: {
agents: ObjectId('SOMETHING')
}
},
{
$lookup: {
from: "agents",
localField: "agents",
foreignField: "_id",
as: "agents"
}
},
{
$lookup: {
from: "agencies",
localField: "agency",
foreignField: "_id",
as: "agency"
}
},
{
$lookup: {
from: "roles",
let: { "roles": "$roles" },
pipeline: [
{ $match: { $expr: { $in: [ "$_id", "$$roles" ] } } },
{
$lookup: {
from:"resources",
let: { "applicants": "$applicants" },
pipeline: [
{ $match: { $expr: { $in: [ "$_id", "$$applicants" ] } } }
],
as: "applicants"
}
},
{
$lookup: {
from: "resources",
let: { "hired": "$hired" },
pipeline: [
{ $match: { $expr: { $in: [ "$_id", "$$hired" ] } } }
],
as: "hired"
}
},
{
$lookup: {
from: "resources",
let: { "rejected": "$rejected" },
pipeline: [
{ $match: { $expr: { $in: [ "$_id", "$$rejected" ] } } }
],
as: "rejected"
}
},
{
$lookup: {
from: "agents",
localField: "hiring_agent",
foreignField: "_id",
as: "hiring_agent"
}
}
],
as: "roles"
}
}
], {
allowDiskUse: true
})
MongoDB is designed to store "boiled" data. It means, in project collection, you need to store redundant information to transform into desired result.
In your case, performance decreases with inner $lookup with let - pipeline for role collection due to this error.
Try to change:
db.collection('projects').aggregate([
{
$match: {
agents: ObjectId('SOMETHING')
}
},
{
$lookup: {
from: "agents",
localField: "agents",
foreignField: "_id",
as: "agents"
}
},
{
$lookup: {
from: "agencies",
localField: "agency",
foreignField: "_id",
as: "agency"
}
},
{
$lookup: {
from: "roles",
let: { "roles": "$roles" },
pipeline: [
{ $match: { $expr: { $in: [ "$_id", "$$roles" ] } } },
{
$lookup: {
from:"resources",
localField: "applicants",
foreignField: "_id",
as: "applicants"
}
},
{
$lookup: {
from: "resources",
localField: "hired",
foreignField: "_id",
as: "hired"
}
},
{
$lookup: {
from: "resources",
localField: "rejected",
foreignField: "_id",
as: "rejected"
}
},
{
$lookup: {
from: "agents",
localField: "hiring_agent",
foreignField: "_id",
as: "hiring_agent"
}
}
],
as: "roles"
}
}
], {
allowDiskUse: true
})
i want use $c.cid
How does this js work?
this customer_url as c and lookup customer use customer_url.cid
{
$lookup: {
from: "customer_url",
localField: "url_id",
foreignField: "url_id",
as: "c"
}
},
{
$unwind: "$c"
},
{
$lookup: {
as: "customer",
from: "customer_dim",
let: {
cid: "$c.cid"
},
pipeline: [
{
$match: {
$expr: {
$eq: ["$_id", "$$cid"]
}
}
}
]
}
}