Find sub-documents using $in with MongoDB - node.js

Question

My task is to find individual authors(comments.user_id) comment on the article (_id)
{
"_id" : ObjectId("56479d9c8510369a4ecea3a9"),
"comments" : [
{
"text" : "222",
"user_id" : ObjectId("563f2db0e2bf6c431b297d45"),
},
{
"text" : "333",
"user_id" : ObjectId("563f2db0e2bf6c431b297d45"),
},
{
"text" : "444",
"user_id" : ObjectId("563f2db0e2bf6c431b297d45"),
},
{
"text" : "55555",
"user_id" : ObjectId("563e3337e2bf6c431b297d41"),
},
{
"text" : "00000",
"user_id" : ObjectId("563f7c0a8db7963420cd5732"),
},
{
"text" : "00001",
"user_id" : ObjectId("563f7c0a8db7963420cd5732"),
}
]
}
My query looks as follows
db.getCollection('messages').find({
'_id': ObjectId("56479d9c8510369a4ecea3a9"),
'comments.user_id': {$in : [
ObjectId("563e3337e2bf6c431b297d41"),
ObjectId("563f7c0a8db7963420cd5732")
]}
})
It returns all comments. Please help to understand why it happens.
Expected Result
{
"_id" : ObjectId("56479d9c8510369a4ecea3a9"),
"comments" : [
{
"text" : "55555",
"user_id" : ObjectId("563e3337e2bf6c431b297d41"),
},
{
"text" : "00000",
"user_id" : ObjectId("563f7c0a8db7963420cd5732"),
},
{
"text" : "00001",
"user_id" : ObjectId("563f7c0a8db7963420cd5732"),
}
]
}
update query (hopelessness)
db.getCollection('messages').find(
{'_id': ObjectId("56479d9c8510369a4ecea3a9")},
{'comments.user_id': {$in: ["563f2db0e2bf6c431b297d45", "563e3337e2bf6c431b297d41"]}},
{'comments.user_id': {$elemMatch: {$in: ["563f2db0e2bf6c431b297d45", "563e3337e2bf6c431b297d41"]}}}
)
db.getCollection('messages').find(
{'_id': ObjectId("56479d9c8510369a4ecea3a9")},
{comments: {$elemMatch: {'user_id': {$in : [ObjectId("563f2db0e2bf6c431b297d45"), ObjectId("563f7c0a8db7963420cd5732")]}}}}
)
I return only 1 record, and I have all the records from these authors

Answer 1

As you've seen, the $ and $elemMatch projection operators only include the first matching element.
To include multiple, filtered array elements in your projection of the comment array, your can use aggregate with the $redact operator instead of find:
db.getCollection('test').aggregate([
{$match: {
'_id': ObjectId("56479d9c8510369a4ecea3a9"),
'comments.user_id': {$in : [
ObjectId("563e3337e2bf6c431b297d41"),
ObjectId("563f7c0a8db7963420cd5732")
]},
}},
{$redact: {
$cond: {
if: {
$or: [
{$eq: ['$user_id', ObjectId("563e3337e2bf6c431b297d41")]},
{$eq: ['$user_id', ObjectId("563f7c0a8db7963420cd5732")]},
{$not: '$user_id'}
]
},
then: '$$DESCEND',
else: '$$PRUNE'
}
}}
])
$redact iterates over each doc like a tree, keeping or trimming the fields of each doc as it's $cond expression dictates.
It gets a bit tricky to wrap your head around $redact, but it's basically saying that if the level's user_id field matches either of the two ObjectIds in your $in, or it's not present (i.e. as it is in the top level of the doc), include the data, otherwise remove it.