Create 4 level nested object aggregate output from a linear mongo collection - mongodb

Question

I have following schema
{
question: String,
answer: Number,
option1: String,
option2: String,
option3: String,
option4: String,
subject: String,
chapter: String,
topic: String,
subtopic: String,
tags: { type: Array, default: [] },
difficulty: String,
media: { type: String, default: "" }
}
I want to create a index from this question schema, where a unique entry in index can be identified by tuple (subject, chapter, topic, subtopic)
{
"subject": {
"chapter": {
"topic": ["subtopics"],
"topic": ["subtopics"],
},
"chapter": {
"topic": ["subtopics"],
"topic": ["subtopics"],
},
},
"subject": {
"chapter": {
"topic": ["subtopics"],
"topic": ["subtopics"],
},
"chapter": {
"topic": ["subtopics"],
"topic": ["subtopics"],
},
},
}
To achieve above result I have been trying some aggregations and so far I am able to get upto
{
"subject": ["chapters"],
"subject": ["chapters"],
"subject": ["chapters"],
"subject": ["chapters"],
}
using following pipeline in mongoose
{
$group: {
_id: "$subject",
chapters: { $push: "$chapter" }
}
},
{
$group: {
_id: null,
chapters: { $push: { "k": "$_id", "v": "$chapters" } }
}
},
{
$replaceRoot: { "newRoot": { "$arrayToObject": "$chapters" } }
}
When I tried to extend my above logic to include topic ans subtopics using
{
$group: {
_id: "$subject",
chapters: { $push: { "k": "$chapter", "v": { topic: "$topic", subtopic: "$subtopic" } } }
}
},
{
$group: {
_id: null,
chapters: { $push: { "k": "$_id", "v": { "$arrayToObject": "$chapters" } } }
}
},
{
$replaceRoot: { "newRoot": { "$arrayToObject": "$chapters" } }
}
I get following output
{
"subject": {
"chapter": {
"topic": "value",
"subtopic": "value"
},
"chapter": {
"topic": "value",
"subtopic": "value"
},
},
"subject": {
"chapter": {
"topic": "value",
"subtopic": "value"
},
"chapter": {
"topic": "value",
"subtopic": "value"
},
},
}
Problem is that chapter is a object of last topic and subtopic only I know where I am making mistake but I don;t know how to fix it.

Answer 1

You can use this aggregation that mainly utilizes $objectToArray, the data restructuring is not so elegant though:
db.collection.aggregate(
[
{
"$group" : {
"_id" : {
"subject" : "$subject",
"chapter" : "$chapter",
"topic" : "$topic"
},
"topics" : {
"$push" : "$subtopic"
}
}
},
{
"$addFields" : {
"topic" : {
"k" : "$_id.topic",
"v" : "$topics"
}
}
},
{
"$group" : {
"_id" : {
"subject" : "$_id.subject",
"chapter" : "$_id.chapter"
},
"topics" : {
"$push" : "$topic"
}
}
},
{
"$addFields" : {
"topics" : {
"$arrayToObject" : "$topics"
}
}
},
{
"$addFields" : {
"chapters" : [
{
"k" : "$_id.chapter",
"v" : "$topics"
}
]
}
},
{
"$addFields" : {
"chapters" : {
"$arrayToObject" : "$chapters"
}
}
},
{
"$group" : {
"_id" : null,
"subject" : {
"$push" : {
"k" : "$_id.subject",
"v" : "$chapters"
}
}
}
},
{
"$project" : {
"final" : {
"$arrayToObject" : "$subject"
}
}
},
{
"$replaceRoot" : {
"newRoot" : "$final"
}
}
]);