CODEWITHSUNDEEP
springmongodbspring-dataspring-data-mongodb
pdeva
pdeva

Reputation: 45491

How to convert this query using Spring Mongodb Aggregation apis?

this is the query I want to perform:

db['applications'].aggregate(
 {$project: {key:1, _id:0}}
 ,{$group: { _id:null, keyList: {$addToSet:"$key"}}} );

This is the code that I am using: 

 Aggregation agg = newAggregation(
            project("key"),
            group("key").addToSet("key").as("keylist")
    );

However, the result produced is this:

Executing aggregation: 
{ "aggregate" : "applications" , "
  pipeline" : [ 
 { "$project" : { "key" : "$key"}} , 
 { "$group" : { "_id" : "$key" , "keylist" : { "$addToSet" : "$key"}}}
]}

Why is the _id key in the group set to the arguments of 'group'. How do I make it null?
How do i get Spring MongoData to produce the query I want?

Upvotes: 2

Views: 5063

Answers (4)

Nabil
Nabil

Reputation: 121

As of the current version of Spring Data MongoDB if you do not specify parameters to group it will group by null.

Aggregation agg = newAggregation(
            project("key"),
            group().addToSet("key").as("keylist")
    );

Upvotes: 1

Robocide
Robocide

Reputation: 6751

Didnt work for me , BUT what did work is to declare in you collection dummy property like this:

Class Model{
...
...
private Integer dummyNotExistsProp
}

and then this is stupid query without matchin and projection but just to demonstrate the solution to the problem: 

Aggregation agg = Aggregation.newAggregation(
                Aggregation.match(Aggregation.group("dummyNotExistsProp").count().as("totalobjs"));

Upvotes: 0

Oliver Drotbohm
Oliver Drotbohm

Reputation: 83081

You can use the method taking a String... and simply pipe in no arguments:

group().addToSet("key").as("keylist")

In the currently released version this renders as:

$group: { _id : {}, keyList : { $addToSet : "$key" }}

While this is not 100% in line with the documentation it does the same as _id : null. We've filed and fixed DATAMONGO-759 which will ship in 1.3.2 and 1.4 M1 of Spring Data Commons.

Upvotes: 4

attish
attish

Reputation: 3150

Unfortunately for the group phase you cannot pass null look at the documentation. 

public static GroupOperation group(String... fields)
Creates a new GroupOperation for the given fields.
Parameters:
fields - must not be null.
Returns:

However you can achive the same with selecting a field name for the group id which is not exists in the collection. Not beautiful but should work.

Aggregation agg = newAggregation(
            project("key"),
            group("ANYSTRINGWHICHISNOTFIELDNAME").addToSet("key").as("keylist")
    );

You do not have to care about the project phase it is good to have smaller data in the pipeline but the result will be the same this way.

Upvotes: 4

Related Questions