Java & MongoDB - Check if field name exists in nested arrays - java

Looking at this simplified example of a record in Mongo:
{
"_id": ObjectId("573b30cb1d62485110330c35"),
"changes": [
{
"after": {
"comments": [
{
"user": {
"idRef": "test1234",
"name": "test user"
}
}
]
}
},
{
"after": {}
}
]
}
Is there a way, using Java's Mongo library, to iterate through each element in each nested array (changes -> after -> comments) and see if the user.idRef field exists?

You can use the following statement
db.collection.find({ "fieldToCheck" : { $exists : true, $ne : null } })
This will return a list of documents matching the criteria.

Related

Java. MongoDB. Sort sub document after unwind

I have users, and skills into the user.
[
{
"_id": "5b91668a0f77e30c11574c88",
"full_name": "John Smith"
"skills": [
{
"_id": "61966603ceb6da478418aac0",
"name": "Java",
"active": true
},
{
"_id": "61966533ceb6da4092707441",
"name": "Scala",
"active": true
},
{
"_id": "619f9dce86b62c370e1c5fcc",
"name": "MongoDB",
"active": true
}
]
}
]
I need to get all user skills in A-Z order by name. I found many solutions, but they are all for MongoDB. I am trying to use the functions of the class com.mongodb.client.model.Agregate, but sorting doesn't work for me (returns in a different order, but it is not clear in what order.) and I do not understand why. Maybe I am do something wrong, or exist another implementation of my task?
var result = getCollection().aggregate(
asList(
match(eq("_id", userId)),
unwind("$skills"),
sort(ascending("skills.name")),
group(null, addToSet("skills", "$skills"))),
Document.class);
Alternatively, you can altogether avoid the $unwind, $sort and $group stages, and use the following aggregation $addFileds (or $project) stage to sort the array elements. This uses the $function aggregate expression (MongoDB v4.4).
Consider an input document:
{
skills: [ { name: "Java" }, { name: "Scala" }, { name: "Mongo" }, { name: "Groovy" } ]
}
The query:
db.collection.aggregate([
{
$addFields: {
skills: {
$function: {
body: function(input) {
const sorter = function(x1, x2) {
if (x1.name > x2.name) {
return 1;
}
else if (x2.name > x1.name) {
return -1;
}
else {
return 0;
};
};
input.sort(sorter);
return input;
},
args: [ "$skills" ],
lang: "js"
}
}
}
}
])
The output with sorted skills array (by name field):
skills: [ { name: "Groovy" }, { name: "Java" }, { name: "Mongo" }, { name: "Scala" } ]

How to get data from mongodb with duplicated parameters?

I have try to create criteria that fetch from data base items.
Here is the code that fetches items from mongo db:
public List<Location> findByListOfId(List<String> locationsIds){
Query query = new Query();
query.addCriteria(Criteria.where("id").in(locationsIds));
return template.find(query, Location.class);
}
here is Location class defenition:
#Document("loaction")
#Data
public class Location {
#Id
private String id;
private long order;
private Date createdAt;
private Date updatedAt;
}
And here is the value of input(List locationsIds) in findByListOfId function:
List<String> locationsIds = {"5d4eee8047206b6d2df212bb","5d4eee8047206b6d2df212bb","5d4eee8047206b6d2df212bb"}
as you can see the input contains the same value three times.
The result that I get from findByListOfId function is a single item with id equal to 5d4eee8047206b6d2df212bb,
while I need to get the numbers of items with the same id as a number of times that exists with in variable(in my case I expect 3 fetched items with id = 5d4eee8047206b6d2df212bb ).
Any idea how this query can be created?
Not sure why you want to do it, but you can do it this way (in Mongo Query Language, you can then translate it in Java).
MongoDB Playground
db.collection.aggregate([
{
$match: {
key: {
$in: [
"5d4eee8047206b6d2df212bb",
"5d4eee8047206b6d2df212bb",
"5d4eee8047206b6d2df212bb"
]
}
}
},
{
"$addFields": {
"itemsArray": [
"5d4eee8047206b6d2df212bb",
"5d4eee8047206b6d2df212bb",
"5d4eee8047206b6d2df212bb"
]
}
},
{
"$unwind": "$itemsArray"
},
])
Using aggregation pipeline, you will add the array as a field using $addFields and then $unwind it (will give you x number of times).
I agree with others it's not something you want to do in production code, but I find the question interesting.
#Yahya's answer works with an assumption that the $match stage returns exactly 1 document.
The more generic pipeline to fetch exact number of documents regardless of how unique the key is and how many duplicates are in the query https://mongoplayground.net/p/546QnaFn4lV :
db.collection.aggregate([
{
$limit: 1
},
{
$project: {
_id: 1,
list: [
"5d4eee8047206b6d2df212bb",
"5d4eee8047206b6d2df212bb",
"6d4eee8047206b6d2df212bc",
"7d4eee8047206b6d2df212bd"
]
}
},
{
"$unwind": "$list"
},
{
"$lookup": {
"from": "collection",
"localField": "list",
"foreignField": "key",
"as": "match"
}
},
{
$project: {
match: {
$cond: [
{
$eq: [
"$match",
[]
]
},
[
{
_id: null,
"key": "$list"
}
],
"$match"
]
}
}
},
{
"$replaceWith": {
$first: "$match"
}
}
])
The first $project passes the list of requested ids to mongo.
The last $project stage returns "null" for requested ids that don't have a matching document.
Here is an aggregate query with required result:
Consider a collection with these documents:
{ _id: 1, a: 11 }
{ _id: 2, a: 22 }
{ _id: 3, a: 99 }
The query in mongo shell with input documents:
var INPUT_IDS = [ 1, 2, 1, 1 ]
db.collection.aggregate([
{
$match: {
_id: { $in: INPUT_IDS }
}
},
{
$group: {
_id: null,
docs: { $push: "$$ROOT" }
}
},
{
$project: {
docs: {
$map: {
input: INPUT_IDS,
as: "inid",
in: {
$let: {
vars: {
matched: {
$filter: {
input: "$docs", as: "doc", cond: { $eq: [ "$$inid", "$$doc._id" ] }
}
}
},
in: { $arrayElemAt: [ "$$matched", 0 ] }
}
}
}
}
}
},
{
$unwind: "$docs"
},
{
$replaceWith: "$docs"
}
])
The output:
{ "_id" : 1, "a" : 11 }
{ "_id" : 2, "a" : 22 }
{ "_id" : 1, "a" : 11 }
{ "_id" : 1, "a" : 11 }

How to write mongo cli query in mongo-template for $in aggregation

This is how my data looks like
{
"_id" : "2011250546437843117",
"name" : "Book",
"textbook" : [
"Maths",
"Science"
],
"language" : [
"English"
],
"isRead" : true,
"isAvailable" : true
}
I have to filter documents based on textbook,and based on that isRead field should be true or false.
my mongo query is
db.user.aggregate([
{
$match: {
"isAvailable": true
}
},
{
$project: {
"textbook": 1,
"name": 1,
"isread": {
$in: [
"Maths",
"$textbook"
]
}
}
}
]);
I have tried to write this using mongo-template
Aggregation aggregation = newAggregation(match(Criteria.where("isAvailable").is(true)),
project("textbook","name"));
I dont understand how to write the $in operator in project stage.
Thankyou in advance.

Spring MongoDB Aggregation Group - Can't get the group query right

I have a document in a MongoDB, which looks like follows.
{
"_id" : ObjectId("5ceb812b3ec6d22cb94c82ca"),
"key" : "KEYCODE001",
"values" : [
{
"classId" : "CLASS_01",
"objects" : [
{
"code" : "DD0001"
},
{
"code" : "DD0010"
}
]
},
{
"classId" : "CLASS_02",
"objects" : [
{
"code" : "AD0001"
}
]
}
]
}
I am interested in getting a result like follows.
{
"classId" : "CLASS_01",
"objects" : [
{
"code" : "DD0001"
},
{
"code" : "DD0010"
}
]
}
To get this, I came up with an aggregation pipeline in Robo 3T, which looks like follows. And it's working as expected.
[
{
$match:{
'key':'KEYCODE001'
}
},
{
"$unwind":{
"path": "$values",
"preserveNullAndEmptyArrays": true
}
},
{
"$unwind":{
"path": "$values.objects",
"preserveNullAndEmptyArrays": true
}
},
{
$match:{
'values.classId':'CLASS_01'
}
},
{
$project:{
'object':'$values.objects',
'classId':'$values.classId'
}
},
{
$group:{
'_id':'$classId',
'objects':{
$push:'$object'
}
}
},
{
$project:{
'_id':0,
'classId':'$_id',
'objects':'$$objects'
}
}
]
Now, when I try to do the same in a SpringBoot application, I can't get it running. I ended up having the error java.lang.IllegalArgumentException: Invalid reference '$complication'!. Following is what I have done in Java so far.
final Aggregation aggregation = newAggregation(
match(Criteria.where("key").is("KEYCODE001")),
unwind("$values", true),
unwind("$values.objects", true),
match(Criteria.where("classId").is("CLASS_01")),
project().and("$values.classId").as("classId").and("$values.objects").as("object"),
group("classId", "objects").push("$object").as("objects").first("$classId").as("_id"),
project().and("$_id").as("classId").and("$objects").as("objects")
);
What am I doing wrong? Upon research, I found that multiple fields in group does not work or something like that (please refer to this question). So, is what I am currently doing even possible in Spring Boot?
After hours of debugging + trial and error, found the following solution to be working.
final Aggregation aggregation = newAggregation(
match(Criteria.where("key").is("KEYCODE001")),
unwind("values", true),
unwind("values.objects", true),
match(Criteria.where("values.classId").is("CLASS_01")),
project().and("values.classId").as("classId").and("values.objects").as("object"),
group(Fields.from(Fields.field("_id", "classId"))).push("object").as("objects"),
project().and("_id").as("classId").and("objects").as("objects")
);
It all boils down to group(Fields.from(Fields.field("_id", "classId"))).push("object").as("objects") that which introduces a org.springframework.data.mongodb.core.aggregation.Fields object that wraps a list of org.springframework.data.mongodb.core.aggregation.Field objects. Within Field, the name of the field and the target could be encapsulated. This resulted in the following pipeline which is a match for the expected.
[
{
"$match" :{
"key" : "KEYCODE001"
}
},
{
"$unwind" :{
"path" : "$values", "preserveNullAndEmptyArrays" : true
}
},
{
"$unwind" :{
"path" : "$values.objects", "preserveNullAndEmptyArrays" : true
}
},
{
"$match" :{
"values.classId" : "CLASS_01"
}
},
{
"$project" :{
"classId" : "$values.classId", "object" : "$values.objects"
}
},
{
"$group" :{
"_id" : "$classId",
"objects" :{
"$push" : "$object"
}
}
},
{
"$project" :{
"classId" : "$_id", "objects" : 1
}
}
]
Additionally, figured that there is no need to using $ sign anywhere and everywhere.

MongoDB Query to match both single entry and array elements

I have a problem with MongoDB QueryBuilder.
Assume I have a number of documents, that can contain one or more users:
{
"_id": "document1",
"data": {
"user": {
"credentials": {
"name": "John",
"lastname": "Watson",
"middle": "Hemish"
}
}
}
}
{
"_id": "document2",
"data": {
"user": [
{
"credentials": {
"name": "John",
"lastname": "Nicholson",
"middle": "Joseph"
}
},
{
"credentials": {
"name": "Mary",
"lastname": "Watson",
"middle": ""
}
}
]
}
}
{
"_id": "document3",
"data": {
"user": [
{
"credentials": {
"name": "John",
"lastname": "Watson",
"middle": "Hemish"
}
},
{
"credentials": {
"name": "John",
"lastname": "Nicholson",
"middle": "Joseph"
}
},
{
"credentials": {
"name": "Mary",
"lastname": "Watson",
"middle": ""
}
}
]
}
}
What I am trying to do is the query, that will return only those documents containing John Watson as a user.
Here what I got so far:
1.
QueryBuilder qb = QueryBuilder.start("credentials.lastname").is("Watson").and("credentials.name").is("John");
DBObject query = QueryBuilder.start("data.user").elemMatch(qb.get()).get();
this query will return only document3: there is no array in document1 and no match in document2 (but I would like it to return document1 and document3)
2.
DBObject query = QueryBuilder.start("data.user.credentials.lastname").is("Watson").and("data.user.credentials.name").is("John").get();
this one will return all three documents: document1 and document3 are desired match, but the query will match as well document2, for it has Watson and John in query fields in the array, no matter that they are separate entries.
Is there any way to make a right query that will return document1 and document3 for John Watson?
I am trying to do it in Java, but any other example would be fine.
Right now I use a workaround combining results from both queries: first I get limit(100) results from the query with elementMatch(), then, if there are less than 100 results, I do the second query and filter all wrong matches. But I hope there is a better and more effective way to get those results.
I could give you at best like the following where user would be in an array as unwind value of the key data. I think a little bit more effort would lead you to the exact format as you want.
I am sharing it as I think it should serve the purpose or anyhow it should help you.
The aggregation query:
db.tuttut.aggregate([
{$unwind:"$data.user"},
{ $project: {
_id:1,
data:1,
temp: {name:"$data.user.credentials.name",
lastname:"$data.user.credentials.lastname"}
} } ,
{ $group:{
_id:"$_id" ,
data: {$addToSet: "$data"} ,
temp:{ $addToSet: "$temp" } } },
{ $match:{ temp:{name:"John",lastname:"Watson"} } } ,
{$project:{_id:1, data:1}}
]).pretty()
Returned Result:
{
"_id" : "document1",
"data" : [
{
"user" : {
"credentials" : {
"name" : "John",
"lastname" : "Watson",
"middle" : "Hemish"
}
}
}
]
}
{
"_id" : "document3",
"data" : [
{
"user" : {
"credentials" : {
"name" : "John",
"lastname" : "Watson",
"middle" : "Hemish"
}
}
},
{
"user" : {
"credentials" : {
"name" : "Mary",
"lastname" : "Watson",
"middle" : ""
}
}
},
{
"user" : {
"credentials" : {
"name" : "John",
"lastname" : "Nicholson",
"middle" : "Joseph"
}
}
}
]
}

Categories