Code Summary: Bucket Pattern

To apply the Bucket pattern to the documents in our bookstore app, we created a bucket document with an id containing the book_id and month timestamp:

{
  "id": {
    "book_id": 1,
    "month": {
      "$date": "2023-11-22T00:00:00.000Z"
    }
  }
}

Next, we added the views array to the bucket document to store all view documents for the month. Each view document included a timestamp and the user_id:

{
  "id": {
    "book_id": 1,
    "month": {
      "$date": "2023-11-22T00:00:00.000Z"
    }
  },
  "views": [
    {
      "user_id": "user00",
      "timestamp": {
        "$date": "2023-11-22T00:00:00.000Z"
      }
    },
    {
      "user_id": "user10",
      "timestamp": {
        "$date": "2023-11-22T00:01:00.000Z"
      }
    },
  ]
}

Although not covered in the previous video, you can also use an aggregation pipeline to implement the Bucket pattern on an existing views collection. For example, imagine you had a views collection where each document corresponds to a user view for a book and consists of a book_id, timestamp and user_id.

{
   "book_id": 34538756,
   "timestamp": {
       "$date": "2023-09-29T08:23:13Z"
   },
   "user_id": 271828
}

In order to retrieve bucket documents containing all view documents of a given book grouped by month and including the total view count, you could use the following aggregation pipeline:

[
   {
       $group: {
           _id: {
               book_id: "$book_id",
               month: {
                   $dateFromParts: {
                       year: { $year: "$timestamp" },
                       month: { $month: "$timestamp" },
                       day: 1,
                   },
               },
           },
           views: {
               $push: {
                   user_id: "$user_id",
                   timestamp: "$timestamp",
               },
           },
       },
   },
   {
       $set: {
           views_count: { $size: "$views" },
       },
   }
]