Storing/Saving only a particular object from list of objects along with parent object in MongoDB - mongodb

Question

The bounty expires in 7 days. Answers to this question are eligible for a +50 reputation bounty.
jarvo69 wants to draw more attention to this question.
In mongodb, I have a master table called category
sample data as below:
{
"_id" : "63d3e01f43aa4e0ee349f841",
"subCategories" : [
{
"subCategoryId" : NumberLong(1),
"name": "Mobile phones"
},
{
"subCategoryId" : NumberLong(2),
"name": "XYZ Machine"
}
]
}
There is another table called product. Sample data as below:
{
"_id" : "63d3e13b43aa4e0ee349f842",
"productId" : NumberLong(1),
"name" : "iphone 14",
"category" : DBRef("category", "63d3e01f43aa4e0ee349f841")
}
While adding new product, only 1 category and 1 subcategory from that selected category can be selected. In my case, I am using #DbRef and I am struggling to find a way through which I can save only 1 subcategory within the product table. Right now it points to an entire object of the category table in which there can be x number of subcategories.
Is it possible to achieve this using #DbRef annotation without changing the database structure and without breaking the category table records in between separate category & subcategory tables ?
May be something like this:
{
"_id" : "63d3e13b43aa4e0ee349f842",
"productId" : NumberLong(1),
"name" : "iphone 14",
"category" : DBRef({"category", "63d3e01f43aa4e0ee349f841"},
"subCategoryId", 1)
}
Using MongoDb version 4+ with Java spring-data-mongo

Answer 1

I don't think it is possible to achieve your expected behaviour without changing the schema. From official doc of DBRef,
DBRefs are a convention for representing a document, rather than a specific reference type.
So DBRef will point to a specific document, instead of certain sub-document array entry.
This leaves us 2 options:
change the category collection to store document like this:
{
"categoryId" : "63d3e01f43aa4e0ee349f841", // this is new
"subCategoryId" : NumberLong(1),
"name": "Mobile phones"
}
Unfortunately this is banned as changing schema is not allowed
add another field in product schema to store the subCategory Id and use it to locate subCategory entries when $lookup
{
"_id": "63d3e13b43aa4e0ee349f842",
"productId": NumberLong(1),
"name": "iphone 14",
"category": {
"$ref": "category",
"$id": "63d3e01f43aa4e0ee349f841"
},
"subCategoryId": NumberLong(1) // this is new
}
the aggregation:
db.product.aggregate([
{
$match: {
"_id": "63d3e13b43aa4e0ee349f842"
}
},
{
"$lookup": {
"from": "category",
"let": {
categoryId: "$category.$id",
subCategoryId: "$subCategoryId"
},
"pipeline": [
{
$match: {
$expr: {
$eq: [
"$$categoryId",
"$_id"
]
}
}
},
{
$unwind: "$subCategories"
},
{
$match: {
$expr: {
$eq: [
"$$subCategoryId",
"$subCategories.subCategoryId"
]
}
}
}
],
"as": "subCategoryLookup"
}
}
])
Mongo Playground
This is also kind of banned as it needs to add one more field to the product schema. But I would still suggest this as this involves a minimal change to the schema.