How to get this parent child query data in Mongo DB - mongodb

Question

I am building a simple messaging system and I have a messageSchema which is setup as below:
const messageSchema = new mongoose.Schema({
replyTo: {
type: mongoose.Schema.ObjectId,
ref: 'Message',
default: null,
},
name: {
type: Object,
required: [true, 'Message has a name'],
},
user: {
type: mongoose.Schema.ObjectId,
ref: 'User',
},
email: {
type: Object,
required: [true, 'Message has a email'],
},
subject: {
type: String,
required: [true, 'Message has a subject'],
},
content: {
type: String,
required: [true, 'Message has content'],
},
createdAt: {
type: Date,
default: Date.now(),
},
status: {
type: String,
default: 'unread',
required: [true, 'Message has a status'],
},});
Sample data below:
[
{
_id: ObjectId('53ed7efca75ca1a5248a281a'),
name: 'Person 1',
createdAt: ISODate('2021-01-01T01:00:00.000Z'),
subject: 'M1',
content: 'M1 content',
replyTo: null,
},
{
_id: ObjectId('53ed80bba75ca1a5248a281b'),
name: 'Person 2',
subject: 'M2 - Reply 1 to M1',
content: 'M2 content',
createdAt: ISODate('2021-01-01T02:00:00.000Z'),
replyTo: ObjectId('53ed7efca75ca1a5248a281a'),
},
{
_id: ObjectId('53ed80bba75ca1a5248a281c'),
name: 'Person 3',
subject: 'M3 - Reply 2 to M1',
content: 'M3 content',
createdAt: ISODate('2021-01-01T03:00:00.000Z'),
replyTo: ObjectId('53ed7efca75ca1a5248a281a'),
},
{
_id: ObjectId('53ed80bba75ca1a5248a281d'),
name: 'Person 4',
subject: 'M4',
content: 'M4 content',
createdAt: ISODate('2021-01-01T02:30:00.000Z'),
replyTo: null,
},
];
I am now trying to query the sample above to produce an inbox style response, so that the root message (no replyTo) is the top-level message, has a latest node with latest message info, and has children(if any) in a children node. See below for desired output.
[
{
_id: ObjectId('53ed7efca75ca1a5248a281a'),
name: 'Person 1',
createdAt: ISODate('2021-01-01T01:00:00.000Z'),
subject: 'M1',
content: 'M1 content',
replyTo: null,
latest: {
_id: ObjectId('53ed80bba75ca1a5248a281c'),
name: 'Person 3',
subject: 'M3 - Reply 2 to M1',
content: 'M3 content',
createdAt: ISODate('2021-01-01T03:00:00.000Z'),
replyTo: ObjectId('53ed7efca75ca1a5248a281a'),
},
children: [
{
_id: ObjectId('53ed80bba75ca1a5248a281b'),
name: 'Person 2',
subject: 'M2 - Reply 1 to M1',
content: 'M2 content',
createdAt: ISODate('2021-01-01T02:00:00.000Z'),
replyTo: ObjectId('53ed7efca75ca1a5248a281a'),
},
{
_id: ObjectId('53ed80bba75ca1a5248a281c'),
name: 'Person 3',
subject: 'M3 - Reply 2 to M1',
content: 'M3 content',
createdAt: ISODate('2021-01-01T03:00:00.000Z'),
replyTo: ObjectId('53ed7efca75ca1a5248a281a'),
},
]
},
{
_id: ObjectId('53ed80bba75ca1a5248a281d'),
name: 'Person 4',
subject: 'M4',
content: 'M4 content',
createdAt: ISODate('2021-01-01T02:30:00.000Z'),
replyTo: null,
latest: {
_id: ObjectId('53ed80bba75ca1a5248a281d'),
name: 'Person 4',
subject: 'M4',
content: 'M4 content',
createdAt: ISODate('2021-01-01T02:30:00.000Z'),
replyTo: null,
},
children: []
},];
Appreciate any help in getting this query sorted. Thanks.

Answer 1

You can use aggregation pipeline stages,
$match replyTo us null condition
$graphLookup to join same collection to get reply messages in children
$addFields to check condition is children is empty then return root document otherwise return last element from children
db.collection.aggregate([
{ $match: { replyTo: null } },
{
"$graphLookup": {
"from": "collection",
"startWith": "$_id",
"connectFromField": "_id",
"connectToField": "replyTo",
"as": "children"
}
},
{
$addFields: {
latest: {
$cond: [
{ $eq: [{ $size: "$children" }, 0] },
"$$ROOT",
{ $arrayElemAt: [{$slice: ["$children", -1]}, 0] }
]
}
}
}
])
Playground
For query optimization you can do last stage $addFields process in your client side.