I have a collection in which the documents sometimes contain a status field and sometimes don't. There is always a data field. I'm trying to form a query to get the latest value for both, but when using the $last operator, I get the values from the latest document and results in status sometimes being null. How can I get it to retrieve the latest defined status value, while still keeping the data value from the latest document?
Current aggregration:
const project = await collection.aggregate([
{
$match: {
projectId: id
}
},
{
$group: {
_id: '$projectId',
status: {
$last: '$status'
},
data: {
$last: '$data'
}
}
}
]).toArray();
You can use $facet and perform multiple query in the parallel on the same set of documents.
db.collection.aggregate([
{
$facet: {
last_status: [
{
"$match": {
status: {
$ne: null
}
}
},
{
"$sort": {
_id: -1
}
},
{
"$limit": 1
}
],
last_data: [
{
"$match": {
data: {
$ne: null
}
}
},
{
"$sort": {
_id: -1
}
},
{
"$limit": 1
}
]
}
},
{
"$project": {
other_fields: {
$first: "$last_data"
},
status: {
$first: "$last_status.status"
}
}
}
])
Working example
Related
I have collection: bookSchema as:
[
{
_id: ObjectId("637d05dc32428ed75ea08d09"),
book_details: {
book_name: "random123",
book_auth: "Amber"
}
},
{
_id: ObjectId("637d0673ce0f17f6c473dee2"),
book_details: {
book_name: "random321",
book_auth: "Amber"
}
},
{
_id: ObjectId("637d069a3d597c8458ebe4ec"),
book_details: {
book_name: "random676",
book_auth: "Amber"
}
},
{
_id: ObjectId("637d06c05b32d503007bcb54"),
book_details: {
book_name: "random999",
book_auth: "Saurav"
}
}
]
Desired O/P to show as:
{
score_ambr: 3,
score_saurabh: 1
}
For this I tried as:
db.bookSchema.aggregate([
{
"$group": {
"_id": {
"$eq": [
"$book_details.book_auth",
"Amber"
]
},
"score_ambr": {
"$sum": 1
}
},
},
{
"$group": {
"_id": {
"$eq": [
"$book_details.book_auth",
"Saurav"
]
},
"score_saurabh": {
"$sum": 1
}
},
}
])
I tried using $group to as I want to group all the matching documents in one and use $count to give the number of count for the matching documents but it doesn't seem to be working and gives the O/P as
O/P:
[
{
"_id": false,
"score_sau": 2
}
]
MongoDB Playground: https://mongoplayground.net/p/cZ64KwAmwlv
I don't know what mean 3 and 1 in your example but if I've understood correctly you can try this query:
The trick here is to use $facet to create "two ways" in the aggregation. One option will filter by Amber and the other one by Saurav.
And then, as values are filtered, you only need yo know the size of the array generated.
db.collection.aggregate([
{
"$facet": {
"score_ambr": [
{
"$match": {
"book_details.book_auth": "Amber"
}
}
],
"score_saurabh": [
{
"$match": {
"book_details.book_auth": "Saurav"
}
}
]
}
},
{
"$project": {
"score_ambr": {
"$size": "$score_ambr"
},
"score_saurabh": {
"$size": "$score_saurabh"
}
}
}
])
Example here
Note that in this way you avoid to use $group.
It looks like what you want is two group twice and create a dynamic key from the book_details.book_auth:
db.bookSchema.aggregate([
{$group: {_id: "$book_details.book_auth", count: {$sum: 1}}},
{$group: {
_id: 0,
data: {$push: {
k: {$concat: ["score_", {$toLower: "$_id"}]},
v: {$sum: "$count"}
}}
}},
{$replaceRoot: {newRoot: {$arrayToObject: "$data"}}}
])
See how it works on the playground example
I am looking for a query for a $match stage in my aggregation which do almost the same, as in this question, but..
if field (named rank in my case) doesn't exists in document, add document to results
but if field, exists, apply $operator condition (in my case it's $max) to this field, and add all documents that suits this condition to the results.
MongoPlayground with example collection.
Result should be like this:
[
{
"method": 3,
"item": 1,
"rank": 3 //because it has field named rank, and suits condition {rank: $max}
},
{
"method": 4,
"item": 1 //we need this, because document doesn't have rank field at all
},
{
"method": 5,
"item": 1 //we need this, because document doesn't have rank field at all
}
]
Things, that I have tried already:
{
$match: {
$or: [
{item: id, rank: {$exists: true, $max: "$rank"}}, //id === 1
{item: id, rank: {$exists: false}} //id === 1
]
}
}
UPD: As for now, probably I don't limit with $match stage only, $project is also relevant after default match, so I could request every document during $match stage by id no matter, have the doc rank field or not, and then, during $project stage do a "separation" by rank $exists
Try this one:
db.collection.aggregate([
{
$match: {
item: id
}
},
{
$group: {
_id: "$item", //<- Change here your searching field
max: {
$max: "$rank" //<- Change here your field to apply $max
},
data: {
$push: "$$ROOT"
}
}
},
{
$unwind: "$data"
},
{
$match: {
$expr: {
$or: [
{
$eq: [
{
$type: "$data.rank"
},
"missing"
]
},
{
$eq: [
"$data.rank",
"$max"
]
}
]
}
}
},
{
$replaceWith: "$data"
}
])
MongoPlayground
I have found an answer, separated from #Valijon's method, but it's also based on the logic above. My query is:
db.collection.aggregate([
{
$match: {
item: id
}
},
{
$project: {
method: 1,
item: 1,
rank: {
$ifNull: [
"$rank",
0
]
}
}
},
{
$group: {
_id: "$item",
data: {
$addToSet: "$$ROOT"
},
min_value: {
$min: "$rank"
},
max_value: {
$max: "$rank"
}
}
},
{
$unwind: "$data"
},
{
$match: {
$or: [
{
$expr: {
$eq: [
"$data.rank",
"$max_value"
]
}
},
{
$expr: {
$eq: [
"$data.rank",
"$min_value"
]
}
},
]
}
}
])
My query is based on $project stage which gives the empty field value 0. It also could be -1, or any value that isn't used in collection. And then I separate results.
MongoPlayground
update so Mohammad Faisal has the best solution.However it breaks when a new document is added lol! so i learned a lot from his code and modified it and it Works! =) the code is all the way in the bottom.
But here's what i said..
So i have this document
{"_id":"5ddea2e44eb407059828d740",
"projectname":"wdym",
"username":"easy",
"likes":0,
"link":["ssss"]
}
{"_id":"5ddea2e44eb407059822d740",
"projectname":"thechosenone",
"username":"easy",
"likes":30,
"link":["ssss"]
}
{"_id":"5ddea2e44eb407059828d740",
"projectname":"thanos",
"username":"wiley",
"likes":10,
"link":["ssss"]
}
and basically what i want is the document that contains the highest
likes with it's associated project name
For example the output would be
"projectname":"thechosenone",
"username":"easy",
"likes":30
}
,
{
"projectname":"thanos",
"username":"wiley",
"likes":10,
}
the code i have for this is the following
db
.collection("projects")
.aggregate([
{
$group: {
_id: { username: "$username" },
likes: { $max: "$likes" }
}
},
{
$project:{projectname:1}
}
])
$project gives me a strange output. However,
the output was correct without the $project.
But i wanted to project the projectname, the user and the highest likes. Thanks for hearing me out :)
heres the solution =)
db
.collection("projects")
.aggregate([
{
$sort: {
likes: -1
}
},
{
$group: {
_id: {
username: "$username"
},
likes: {
$max: "$likes"
},
projectname: {
$push: "$projectname"
},
link: {
$push: "$link"
}
}
},
{
$project: {
username: "$_id.username",
projectname: {
$arrayElemAt: ["$projectname", 0]
},
link: {
$arrayElemAt: ["$link", 0]
}
}
}
])
.toArray()
If you don't have to use $group this will solve your problem:
db.projects.aggregate([
{$sort:{likes:-1}},
{$limit:1}
]).pretty()
the result would be
{
"_id" : ObjectId("5ddee7f63cee7cdf247059db"),
"projectname" : "thechosenone",
"username" : "easy",
"likes" : 30,
"links" : ["ssss"]
}
Try this:-
db.collection("projects").aggregate([
{
$group: {
_id: { username: "$username" },
likes: { $max: "$likes" },
projectname: { $push : { $cond: [ { $max: "$likes" }, "$projectname", "" ]}}
}
}
,
{
$project:{
username:"$_id.username",
projectname:{"$reduce": {
"input": "$projectname",
"initialValue": { "$arrayElemAt": ["$projectname", 0] },
"in": { "$cond": [{ "$ne": ["$$this", ""] }, "$$this", "$$value"] }
}},
likes:1
}
}
])
data=[{
locId: '332wn',
locadetails: [
{ loc: 'ny',
status: true
},
{ loc: 'ca',
status: null
},
{ loc: 'tx',
status: null
}
]
}]
I have following query that is trying to find all the locdetails that have open value as null or false
Loc.find({'locId': id}, {'locadetails' : {$elemMatch: {'status': {$ne: true}}}}, (err, locs)=>{
if(err) {
retrun callback(err);
}
callback(null, locs)
});
Problem I have is this query will only return one value o locadetails with null while it should return two as seen in the data.
Please let me know what to do so I can get whole array of items that have status field as null or false ...Thanks
$elemMatch will return first matching element from an array based on a condition. Use Aggregation instead.
Both the $ operator and the $elemMatch operator project the first
matching element from an array based on a condition. Reference
Loc.aggregate([
{ $match: { "locId": "332wn" } },
{ $unwind: "$locadetails" },
{ $match: { "locadetails.status": { $ne: true } } },
{ $group: { _id: "$_id", locId: { $first: "$locId" }, locadetails: { $push: "$locadetails" }, } }
])
on my project i have users that complete combinations (called sessions) of courses. the fact of playing a course is called an attempt. During the attempt they can close it and come back later (so we keep a timelog object).
I have a request from the client which needs to return for each session, the users (and their attempts) that have played whole or part of their session during a certain timeframe.
During a certain timeframe means that the client sends a begin and end date and we count a user for a specific session if:
- the first attempt has begun before the end of the timeframe => the started of the first timelog of the first < ending date
- the last attempt has been finished after the begining of the timeframe => the end of the last timelog of the last attempt > starting date
Here is an example of an attempt object (the only one we need to use here):
{
"_id" : ObjectId("5b9148650ab5f43b5e829a4b"),
"index" : 0,
"author" : ObjectId("5acde2646055980a84914b6b"),
"timelog" : [
{
"started" : ISODate("2018-09-06T15:31:49.163Z"),
"ended" : ISODate("2018-09-06T15:32:03.935Z")
},
...
],
"session" : ObjectId("5b911d31e58dc13ab7586f9b")}
My idea was to make an aggregate on the attempts, to group those using author and session as an _id for the $group stage, and to push all the attempts of the user for this particular session into an array userAttempts.
Then to make an $addField stage to retrieve the started field of the first timelog of the first attempt and the last ended of the last attempt.
And finally to $filter or $match using those new fields.
Here is my aggregate:
const newDate = new Date()
_db.attempts.aggregate([
{ $match: {
author: { $in: programSessionsData.users },
$or: [{ programSession: { $in: programSessionIds } }, { oldTryFor: { $in: programSessionIds } }],
globalTime: $ex,
timelog: $ex }
},
{
$group: {
_id: {
user: "$author",
programSession: "$programSession"
},
userAttempts: { $push: { attemptId: "$_id", lastTimelog: { $arrayElemAt: ["$timelog", -1] }, timelog: "$timelog" } }
}
},
{
$addFields: { begin: { $reduce: {
input: "$userAttempts",
initialValue: newDate,
in: {
$cond: {
if: { $lt: ["$$this.timelog.0.started", "$$value"] },
then: "$$this.timelog.0.started",
else: "$$value"
} }
} } }
}
I also tried this for the addFields stage:
{
$addFields: { begin: { $reduce: {
input: "$userAttempts",
initialValue: newDate,
in: { $min: ["$$this.timelog.0.started", "$$value] }
} } }
}
However everytime begin is an empty array.
I do not really know how i can extract those two date, or compare dates between them.
To Note: the end one is more difficult that is why i have to first extract lastTimelog. If you an other method i would gladly take it.
Also this code is on a node server so i cannot use ISODate. and the mongo version used is 3.6.3.
After playing with aggregate a bit i came up with 2 solutions:
Solution 1
_db.attempts.aggregate([
{ $match: {
query
},
{
$group: {
_id: {
user: "$author",
programSession: "$programSession"
},
userAttempts: { $push: { attemptId: "$_id", timelog: "$timelog" } }
}
}, {
$addFields: {
begin: { $reduce: {
input: "$userAttempts",
initialValue: newDate,
in: { $min: [{ $reduce: {
input: "$$this.timelog",
initialValue: newDate,
in: { $min: ["$$this.started", "$$value"] }
} }, "$$value"] }
} },
end: { $reduce: {
input: "$userAttempts",
initialValue: oldDate,
in: { $max: [{ $reduce: {
input: "$$this.timelog",
initialValue: oldDate,
in: { $max: ["$$this.ended", "$$value"] }
} }, "$$value"] }
} }
}
},
{
$match: {
begin: { $lt: req.body.ended },
end: { $gt: req.body.started }
}
}
], { allowDiskUse: true });
newDate is today and oldDate is an arbitrary date in the past.
I had to chain 2 reduce because "$$this.timelog.0.started" would always return nothing. Don't really know why though.
Solution 2
_db.attempts.aggregate([
{ $match: {
query
},
{
$addFields: {
firstTimelog: { $arrayElemAt: ["$timelog", 0] },
lastTimelog: { $arrayElemAt: ["$timelog", -1] }
}
},
{
$group: {
_id: {
user: "$author",
programSession: "$programSession"
},
begin: { $min: "$firstTimelog.started" },
end: { $max: "$lastTimelog.ended" },
userAttempts: { $push: { attemptId: "$_id", timelog: "$timelog"} }
}
},
{
$match: {
begin: { $lt: req.body.ended },
end: { $gt: req.body.started }
}
}
], { allowDiskUse: true });
This one is a lot more straight forward and seems simpler, but oddly enough, from my testing, Solution 1 is always quicker at least in the object distribution for my project.