mongoDB 聚合 $lookup 并从现有数组中推送一些信息

Question

I have many collections in my MongoDB database. Aggregation work's good but I can't push some fields to the output I needed,

collection A is:

{
  _id: some mongodbID
  //...fields
  items: [
   {
     _id: someId,
     color: someId <---- Im aggregate this with lookup
     neededFieldToPush: 123
   },
   {
     _id: someId,
     color: someId <---- Im aggregate this with lookup
     neededFieldToPush: 566
   }
  ]
}

my query is:

await Invoice.aggregate([
  { $match: query },
  { $unwind: "$items" },
  //colors
  {
    $lookup: {
      from: "colors",
      localField: "items.itemColor",
      foreignField: "_id",
      as: "itemColor"
    }
   },
   {
     $addFields: {
       "prMove.itemColor": { $arrayElemAt: ["$itemColor.colorName", 0] },
     }
    },
    {
      $group: {
        _id: "$_id",
        items: { $push: "$items" }, <-- original items
        prMove: { $push: "$prMove" },
       }
     },
   ])
   .sort({date: -1})
   .skip(+req.query.offset)
   .limit(+req.query.limit)

I need to have output like this:

_id: someId,
items: [//original items],
prMove: [
  {
    itemColor: some color name, <--- it's works fine
    neededFieldToPush: 123
  },
  {
    itemColor: some color name, <--- it's works fine
    neededFieldToPush: 566
  },      
]

so, how I can push neededFieldToPush field into prMove object?

thank you

Answer 1

The implementation of your query is expensive because you have used $unwind stage and $group stage, it will impact performance,

The second thing is the $sort , $skip and $limit functions will not work on the aggregate function of mongoose, you have use stages for that,

The third thing is to use sort, $skip and $limit stages immediately after $match stage, so it will improve performance while you create an index on required fields.

Improved query,

• $match your query
• $sort sort stage
• $skip stage to pass offset
• $limit stage to limit documents
• $lookup with colors collection and pass items.color as localField
• $addFields to edit items array with color name in items
• $map to iterate loop of items array
• $reduce to iterate loop of itemColor array result from lookup, check condition if _id match then get color name and set value in color field
• $mergeObjects to merge new color field and current fields of the item object
• $$REMOVE to remove itemColor array because it is not needed now

let offset = req.query.offset;
let limit = req.query.limit;
await Invoice.aggregate([
  { $match: query },
  { $sort: { date: -1 } },
  { $skip: offset },
  { $limit: limit },
  {
    $lookup: {
      from: "colors",
      localField: "items.color",
      foreignField: "_id",
      as: "itemColor"
    }
  },
  {
    $addFields: {
      items: {
        $map: {
          input: "$items",
          as: "i",
          in: {
            $mergeObjects: [
              "$$i",
              {
                color: {
                  $reduce: {
                    input: "$itemColor",
                    initialValue: "",
                    in: {
                      $cond: [
                        { $eq: ["$$this._id", "$$i._id"] },
                        "$$this.colorName",
                        "$$value"
                      ]
                    }
                  }
                }
              }
            ]
          }
        }
      },
      itemColor: "$$REMOVE"
    }
  }
])

Playground