Aggregate function on mongo running very slow when running query - mongodb

Question

Tried running a query on the aggregate function on Mongo which is currently taking 16 seconds when the result i wished for was under a second
{
"$lookup": {
"from": "session_attendances",
"let": { "id": "$_id" },
"pipeline": [
{
"$match": {
"$expr": {
"$eq": ["$student", "$$id"]
}
}
},
{
"$project": {
"attendance_code": "$attendance_code"
}
}
],
"as": "attendance"
}
},
{
// keep only matched students, can skip this and modifiy the next phase incase no such documents exists.
"$unwind": "$attendance"
},
{
"$lookup": {
"from": "attendance_codes",
"let": { "attendance_code": "$attendance.attendance_code" },
"pipeline": [
{
"$project": {
"type": 1
}
},
{
"$match": {
"$expr": {
"$eq": ["$_id", "$$attendance_code"]
}
}
}
],
"as": "attendance_code"
}
},
{
//again assuming we want to keep matched docs otherwise why lookup?
"$unwind": "$attendance_code"
},
{
"$group": {
"_id": { "a": "$attendance.attendance_code", "id": "$_id" },
"total": { "$sum": 1 },
"data": { "$first": "$$ROOT" } // if u want to keep document data
}
}
Hoping that some one can give me an answer to which part of my code is making the run time so slow.

Answer 1

Its not clear what your end goal is, if you wish to clarify that it would help me give an alternative to your current aggregation
With that said the second lookup stage is "useless" as you group right after without using any of the data gained by it, removing it will still get you the exact same result and save some time.
Assuming the second lookup is needed for some reason i recommend not nesting it but rather use after the first one, like so:
{
$lookup: {
from: 'session_attendances',
let: { 'id': '$_id' },
pipeline: [
{
"$match": {
"$expr": {
"$eq": ["$student", "$$id"]
}
}
}
,{
$project: {
attendance_code: '$attendance_code'
}
}
],
as: 'attendance'
}
},
{// keep only matched students, can skip this and modifiy the next phase incase no such documents exists.
$unwind: "$attendance"
},
{
$lookup: {
from: 'attendance_codes',
let: { 'attendance_code': '$attendance.attendance_code' },
pipeline: [
{
$project: {
type: 1
}
},
{
"$match": {
"$expr": {
"$eq": ["$_id", "$$attendance_code"]
}
}],
as: 'attendance_code'
}
}
},
{ //again assuming we want to keep matched docs otherwise why lookup?
$unwind: "$attendance_code"
},
{
$group: {
_id: {a: "$attendance.attendance_code", id: "$_id"}
total: { $sum: 1 },
data: {$first: "$$ROOT"} // if u want to keep document data
}
}
This should give you better performance, i also recommend dropping the projects stages, unless the documents are very large this usually does not end up helping performance but actually hurting it.