Reputation: 684
I'm building a new LMS in mongodb and I have the following collections:
Courses
{
"_id": ObjectId("5f6a6b159de1304fb885b194"),
"title": "Course test",
"sections": [
{
"_id": ObjectId("5f6a6b159de1304fb885b195"),
"title": "Section 1 - introduction",
"order": 1,
"modules": [
{
"_id": ObjectId("5f6a6b159de1304fb885b196"),
"module_FK_id": ObjectId("5f6a6b149de1304fb885b135"),
"title": "Module 1",
"order": 1
},
{
"_id": ObjectId("5f6a6b159de1304fb885b198"),
"module_FK_id": ObjectId("5f6a6b149de1304fb885b14a"),
"title": "Module 2",
"order": 2
},
]
},
{
"_id": ObjectId("5f6a6b149de1304fb885b175"),
"title": "Section 2 - How to do something",
"order": 2,
"modules": [
{
"_id": ObjectId("5f6a6b149de1304fb885b141"),
"module_FK_id": ObjectId("5f6a6b149de1304fb885b150"),
"title": "Module 1",
"order": 1
},
{
"_id": ObjectId("5f6a6b149de1304fb885b15f"),
"module_FK_id": ObjectId("5f6a6b149de1304fb885b18e"),
"title": "Module 2",
"order": 2
},
]
},
]
}
Modules (only one as example)
{
"_id": ObjectId("5f6a6b149de1304fb885b135"),
"text": "Lorem ipsum...",
"mediaUrl": "urllinkhere"
}
As shown, I choose to have embedded documents for sections and modules titles but I need also a second collection, modules, because each module contains a large amount of text and my course document may get too big quickly.
Now I need to rebuild the entire document as if it was completely embedded. Here's an example:
{
"_id": ObjectId("5f6a6b159de1304fb885b194"),
"title": "Course test",
"sections": [
{
"_id": ObjectId("5f6a6b159de1304fb885b195"),
"title": "Section 1 - introduction",
"order": 1,
"modules": [
{
"_id": ObjectId("5f6a6b159de1304fb885b196"),
"module_FK_id": ObjectId("5f6a6b149de1304fb885b135"),
"title": "Module 1",
"order": 1
"text": "Lorem ipsum...",
"mediaUrl": "urllinkhere"
},
// last two fields from collection "modules"
I'm trying different combination of aggregation and lookups but I can't obtain the desired result.
Can anybody help me?
Upvotes: 0
Views: 195
Reputation: 1590
You can construct the aggregation pipeline like this below, just remember to
$group in the reverse order of
$unwind operations
db.courses.aggregate([
{
$unwind: "$sections"
},
{
$unwind: "$sections.modules"
},
{
"$lookup": {
"from": "modules",
"localField": "sections.modules.module_FK_id",
"foreignField": "_id",
"as": "module_details"
}
},
{
$unwind: {
path: "$module_details",
preserveNullAndEmptyArrays: true
}
},
{
"$project": {
title: 1,
sections: {
_id: "$sections._id",
modules: {
_id: "$sections.modules._id",
module_FK_id: "$sections.modules.module_FK_id",
order: "$sections.modules.order",
title: "$sections.modules.title",
mediaUrl: "$module_details.mediaUrl",
text: "$module_details.text"
},
order: "$sections.order",
title: "$sections.title"
}
}
},
{
"$group": {
"_id": "$sections._id",
main_id: {
$first: "$_id"
},
main_title: {
$first: "$title"
},
order: {
$first: "$sections.order"
},
title: {
$first: "$sections.title"
},
modules: {
$push: "$sections.modules"
}
}
},
{
"$project": {
"_id": "$main_id",
"title": "$main_title",
section: {
_id: "$_id",
modules: "$modules",
title: "$title",
order: "$order",
}
}
},
{
$group: {
_id: "$_id",
title: {
$first: "$title"
},
sections: {
$push: "$section"
}
}
}
])
Upvotes: 1