MongoDB aggregation grouping fill missing values - mongodb

Question

I'm using MongoDB aggregation framework. I have a Mongo collection with documents like this:
{
'step': 1,
'name': 'house',
'score': 2
}
{
'step': 1,
'name': 'car',
'score': 3
}
{
'step': 2,
'name': 'house',
'score': 4
}
I'm grouping the documents with same 'step' and pushing 'name' and 'score' into an array of objects. What I get is:
{
'step': 1,
'scores':
[
{'name':'house','score':2},
{'name':'car','score':3}
]
}
{
'step': 2,
'scores':
[
{'name':'house','score':4}
]
}
For each 'step' I need to copy the value of previous 'step' in case that a 'name' does not exists. I should have something like this:
{
'step': 1,
'scores':
[
{'name':'house','score':2},
{'name':'car','score':3}
]
}
{
'step': 2,
'scores':
[
{'name':'house','score':4},
**{'name': 'car', 'score':3}**
]
}
At the second document the element {'name':'car','score':3} has been copied from the previous document because at 'step:2' there is not documents having 'score' for 'car'.
I'm not able to figure out how to do this operation with MongoDB aggregation. Some help will be very appreciated.

Answer 1

Required to use $lookup with pipeline, look below step by step,
$group by step and push all scores in one array scores
push all name in names of each score of particular step, we will use in match condition inside lookup
db.collection.aggregate([
{
$group: {
_id: "$step",
scores: {
$push: {
name: "$name",
score: "$score"
}
},
names: { $push: "$name" }
}
},
$unwind scores because its array and we are going to lookup
{ $unwind: "$scores" },
$lookup let variables step(_id) and names for pipeline level
$match condition with expression $expr there are 3 conditions
check the size of names It should be one(1), either its car or house,
match step number, it should be equal
match not in for ex. if car is already available then it will search for house in lookup, need to use separate $not and than $in
$project to show required fields
lookup result will store in clone_score
{
$lookup: {
from: "collection",
let: {
step_id: { $subtract: ["$_id", 1] },
names: "$names"
},
pipeline: [
{
$match: {
$expr: {
$and: [
{ $eq: [{ $size: "$$names" }, 1] },
{ $eq: ["$$step_id", "$step"] },
{ $not: [{ $in: ["$name", "$$names"] }] }
]
}
}
},
{
$project: {
_id: 0,
name: 1,
score: 1
}
}
],
as: "clone_score"
}
},
$group by step(_id) and push all scores in one array scores, keep first clone_score
{
$group: {
_id: "$_id",
scores: { $push: "$scores" },
clone_score: { $first: "$clone_score" }
}
},
from above pipelines, we have two separate array scores and clone_score now,
$project we need to concat both of them in scores
{
$project: {
_id: 0,
step: "$_id",
scores: {
$concatArrays: ["$scores", "$clone_score"]
}
}
}
])
Playground: https://mongoplayground.net/p/Fytf7NEU7uG