如何根据字符串的引用数组是否有共同的项目来查找包含字符串数组的文档？

Question

Given a reference array of strings: ["Comedy", "Horror", "Romance"] , I would like to query a Movie model with this schema:

const MovieSchema = new Schema({
    _id: {
        type: Types.ObjectId,
        required: true
    },
    title: {
        type: String,
        required: true
    },
    categories: [{ type: String }],
});

Such that I will get results where I will get Movie s with categories in common with the reference array, sorted by the amount of elements it has in common with the reference array. For example:

[
    {
        _id: "57",
        title: "Sample Movie A",
        categories: ["Comedy", "Horror", "Romance", "Family"]
    },
    {
        _id: "92",
        title: "Sample Movie B",
        categories: ["Comedy", "Romance", "Family", "Coming of Age"]
    }
]

Note that Movie A is the first in the results because it has 3 items in common with the reference array while Movie B only has 2 items in common.

How can I achieve this query using Mongoose 5.11.16?

Answer 1

You could use $setIntersection to get a count of matching elements, add the resulting array-size as a new field to each document and and then sort by this field.

You could then extend the query to filter matches with a count greater than 0 and remove the categoryCount from the output, eg

Movie.aggregate([
  {
    "$addFields": {
      "categoryCount": {
        $size: {
          $setIntersection: [
            [
              "Comedy",
              "Horror",
              "Romance"
            ],
            "$categories"
          ]
        }
      }
    }
  },
  {
    "$match": {
      categoryCount: {
        $gt: 0
      }
    }
  },
  {
    "$sort": {
      categoryCount: -1
    }
  },
  {
    "$project": {
      categoryCount: 0
    }
  }
])

Example on mongoplayground: https://mongoplayground.net/p/ZlUNfB82FRK