$lookup and replace array of ids with values - mongodb

Question

Consider below Mongo DB 3.2 Document,
Clients Collection Document
{
"_id" : "gcJk4eRRo2WCbgWSL",
"services" : [
"2tLX8ALYfRvbgiurZ",
"wfE5MqgHfu9QKtK7d",
"MEZtABSEeskuRivXJ"
]
}
Also, Services Collection Documents
{ "_id" : "2tLX8ALYfRvbgiurZ", "name" : "GSTR 1" }
{ "_id" : "wfE5MqgHfu9QKtK7d", "name" : "GSTR 2" }
{ "_id" : "MEZtABSEeskuRivXJ", "name" : "GSTR 3" }
Now, the values in services array field in Clients is associated to _id of Services Collection.
Below is the code that I am currently executing,
db.getCollection('Clients').aggregate(
[
{ "$unwind" : { path: "$services"}},
{
"$lookup" : { from: "Services", localField: "services", foreignField: "_id", as: "services" }
},
{ "$unwind" : { path: "$services", preserveNullAndEmptyArrays: true }},
{ "$project" : {
_id : 1, services: '$services.name'
}
}
]
);
Output of Above code execution is,
{ "_id" : "gcJk4eRRo2WCbgWSL", "services" : "GSTR 1" }
{ "_id" : "gcJk4eRRo2WCbgWSL", "services" : "GSTR 2" }
{ "_id" : "gcJk4eRRo2WCbgWSL", "services" : "GSTR 3" }
But Expected output is as below,
{
"_id" : "gcJk4eRRo2WCbgWSL",
"services" : "GSTR 1, GSTR 2, GSTR 3"
}
Any help is highly appreciated.

Answer 1

You can add additional $group by your _id with $push to merge your services into one array.
db.Clients.aggregate(
[
{ "$unwind" : { path: "$services"} },
{
"$lookup" : { from: "Services", localField: "services", foreignField: "_id", as: "services" }
},
{ "$unwind" : { path: "$services", preserveNullAndEmptyArrays: true }},
{ "$project" : {
_id : 1, services: '$services.name'
}
},
{
"$group": {
_id: "$_id",
"services": { "$push": "$services"}
}
}
]
);

Answer 2

Here's how you do that - there is no need for any $unwinds as long as you do not care about the item order inside your string.
db.getCollection('Clients').aggregate([
{
"$lookup": {
from: "Services",
localField: "services",
foreignField: "_id",
as: "services"
}
}, {
"$project": {
"_id" : 1,
"services": {
$substr:
[
{
$reduce: { // transform array of strings into concatenated string
input: '$services.name',
initialValue: '',
in: {
$concat: ['$$value', ', ', '$$this']
}
}
},
2, // remove first two characters as they will be ', '
-1
]
}
}
}])
This, however, can potentially return something like the following document (note the order of the entries in the string):
{
"_id" : "gcJk4eRRo2WCbgWSL",
"services" : "GSTR 1, GSTR 3, GSTR 2"
}
If you need the items inside the string to be sorted you can do it this way:
db.getCollection('Clients').aggregate([
{
"$lookup": {
from: "Services",
localField: "services",
foreignField: "_id",
as: "services"
}
}, {
$unwind: "$services" // flatten the services array
}, {
$sort: {
"services.name": 1 // sort all documents by service name
}
}, {
$group: { // group the everything back into the original structure again
"_id": "$_id", // we want one group per document id
"services": {
$push: "$services" // and all its services in an array - this time reliably sorted!
}
}
}, {
"$project": {
"_id": 1,
"services": {
$substr:
[
{
$reduce: { // transform array of strings into concatenated string
input: '$services.name',
initialValue: '',
in: {
$concat: ['$$value', ', ', '$$this']
}
}
},
2, // remove first two characters as they will be ', '
-1
]
}
}
}])
This way you reliably get the entries sorted by name:
{
"_id" : "gcJk4eRRo2WCbgWSL",
"services" : "GSTR 1, GSTR 2, GSTR 3"
}

Answer 3

You can try using concatArrays.
Like what mickl answered, you can do grouping first and project it to concat the array value.
//previous code
{
"$group": {
_id: "$_id",
"services": { "$push": "$services"}
}
},{ $project: {
_id : "$_id",
"services": { $concatArrays: [ "$services" ] } }
}
hope it helps..