given this fantasy dataset (sorry in advance but I couldn't manage to format it properly):

I would need to make a MongoDb pipe where a new field is created and, this new field, should be able to add a parameter valued with an empty string whene the "k" parameter is missing or empty.

Here is my attempt:

...    
{
    "$addFields": {
        "colors_field": {
            "r": "$colors.r",
            "g": "$colors.g",
            "b": "$colors.b",
            "k": {
                "$cond": {
                    "if": {
                       "$or": [
                            {"$eq": [ "$colors.k", "" ]},
                            {"$eq": [ "$colors.k", null ]},
                            {"$colors.k": { "$exists": false}}
                            ]
                    },
                    "then": "",
                    "else": "$colors.k"
                }
            }
        }
    }
}

I tried with the $exists but I can't make it work because of the way I call the value to check: either I put "$colors.k" and it returns a "MongoError: Unrecognized expression" or I don't and it'll return a "MongoError: FieldPath field names may not contain '.'". I also tried to check for the length of that field but it'll crash if said field is missing.

CodePudding user response：

The problem with your query is that you are using a query operatos $exists inside a pipeline (only $match stage allows this).

We dont have an "$exists" operator in aggregation but we can do it using $type and "missing", but here you want exists or null, so $ifnull is the right way to do it(that does exaclty this,name is missleading, its if not-exists or null).

Query (3 ways to do it)

1. $type and "missing" (the right way to go if you wanted only exists), here you want null, also so it becomes bigger query
2. $ifnull this is the shortest way
3. $cond not-exists or null are both false values, so this works also, BUT be careful with this, because if color was false, it would make it also ""

*in your case the short and safe solution is the 2

Test code here

db.collection.aggregate([
  {
    "$set": {
      "k-cond": {
        "$cond": [
          {
            "$or": [
              {
                "$eq": [
                  "$colors.k",
                  null
                ]
              },
              {
                "$eq": [
                  {
                    "$type": "$colors.k"
                  },
                  "missing"
                ]
              }
            ]
          },
          "",
          "$colors.k"
        ]
      },
      "k-ifnull": {
        "$ifNull": [
          "$colors.k",
          ""
        ]
      },
      "k-if": {
        "$cond": [
          "$colors.k",
          "$colors.k",
          ""
        ]
      }
    }
  }
])