Upsert field with aggregation

Working with Data
,
1

This is a continuation of this thread but it got locked.

I marked the solution that works with seasons arrays. However, I also got a solution from @Prasad_Saya that works with seasons objects (which is the desired solution)

My collection has objects that are like this:

{
    totalWins: 12,
    seasons: {
        1: {
            wins: 10
        },
        2: {
            wins: 2
        }
    }
}

I want to sum the wins from the seasons objects and put that sum into the field totalWins . However, it is possible that the seasons field might not exist, or it has no objects in itself, in which case totalWins should be 0.

The aggregation I got is:

db.test.aggregate([
  { 
      $project: { 
          seasons: { "$objectToArray": "$seasons" } 
      } 
  },
  { 
      $project: { 
          totalWins: { 
              $reduce: { 
                  input: { $ifNull: [ "$seasons", [] ] }, 
                  initialValue: 0, 
                  in: { $add: [ "$$value", "$$this.v.wins" ] }
              }
          }
      }
  }
])

This returns a collection with totalWins in the objects. However, I want the totalWins to be updated/inserted to the original collection so I did updateMany along with the aggregation. The problem is that doing so causes the original documents to be replaced by the documents returned by the aggregation. As a result, I lose all of the other fields that were in the original documents before, and now my documents only have _id and totalWins. So my question is how do I run the aggregation, take the field totalWins and append it to the existing documents without overwriting them?

2

You don’t need to $project anything.

Just use a $set: { totalWins: {expression}} and your aggregated document will emerge from the $set stage with an extra, added field.

1 Like
3

As @Jack_Woehr had mentioned, for the update do use $set. For the query
use $addFields instead of $project in the aggregation - this will ensure that the existing fields will remain in addition to the updated/created fields.

1 Like
4 
db.test.updateMany({}, [
  { 
      $project: { 
          seasons: { "$objectToArray": "$seasons" } 
      } 
  },
  { 
      $set: { 
          totalWins: { 
              $reduce: { 
                  input: { $ifNull: [ "$seasons", [] ] }, 
                  initialValue: 0, 
                  in: { $add: [ "$$value", "$$this.v.wins" ] }
              }
          }
      }
  }
])

Do you mean like this? If so this didn’t work. Fields in the original documents were gone, now the documents only have _id, seasons as an array instead of an object, and totalWins. Am I misunderstanding something here?

5

Just replace the $project with $addFields in the first stage of the aggregation.

6

Doing so mutates the seasons field, but the other fields are fine. Instead of ‘seasons’ originally being an object, now it’s an array. I want to preserve to original ‘seasons’ object as well.

7

This will preserve the seasons:

seasons_array: { "$objectToArray": "$seasons" }

Then, use that in the following reduce operation:

input: { $ifNull: [ "$seasons_array", [] ] }

8

Doing so results in both seasons and seasons_array in each document. So I ended up adding $unset: seasons_array to remove the field.

Now my question is: is this the best way to approach this? Or is there like another way to sort of declare a local variable to store the temporary array so that it doesn’t get added to the document in the end.

9

Yes, thats the correct way of doing it.

2 Likes
10

This topic was automatically closed 5 days after the last reply. New replies are no longer allowed.