Aggregate Data in MongoDB Atlas - Functions

On this page

Overview

Data Model
Run an Aggregation Pipeline
Find Data with Atlas Search
Aggregation Stages
Filter Documents
Group Documents
Project Document Fields
Add Fields to Documents
Unwind Array Values

Atlas Device Sync, Atlas Edge Server, Data API, and HTTPS Endpoints are deprecated. Refer to the deprecation page for details.

Overview

The examples on this page demonstrate how to use the MongoDB Query API in a function to aggregate documents in your Atlas cluster.

MongoDB aggregation pipelines run all documents in a collection through a series of data aggregation stages that allow you to filter and shape documents as well as collect summary data about groups of related documents.

Note

Supported Aggregation Stages

Atlas App Services supports nearly all MongoDB aggregation pipeline stages and operators, but some stages and operators must be executed within a system function. See Aggregation Framework Limitations for more information.

Data Model

The examples on this page use a collection named store.purchases that contains information about historical item sales in an online store. Each document contains a list of the purchased items, including the item name and the purchased quantity, as well as a unique ID value for the customer that purchased the items.

{
  "title": "Purchase",
  "required": ["_id", "customerId", "items"],
  "properties": {
    "_id": { "bsonType": "objectId" },
    "customerId": { "bsonType": "objectId" },
    "items": {
      "bsonType": "array",
      "items": {
        "bsonType": "object",
        "required": ["name", "quantity"],
        "properties": {
          "name": { "bsonType": "string" },
          "quantity": { "bsonType": "int" }
        }
      }
    }
  }
}

Snippet Setup

To use a code snippet in a function, you must first instantiate a MongoDB collection handle:

exports = function() {
  const mongodb = context.services.get("mongodb-atlas");
  const itemsCollection = mongodb.db("store").collection("items");
  const purchasesCollection = mongodb.db("store").collection("purchases");
  // ... paste snippet here ...
}

Run an Aggregation Pipeline

You can execute an aggregation pipeline using the collection.aggregate() method.

The following function snippet groups all documents in the purchases collection by their customerId value and aggregates a count of the number of items each customer purchases as well as the total number of purchases that they made. After grouping the documents the pipeline adds a new field that calculates the average number of items each customer purchases at a time, averageNumItemsPurchased, to each customer's document:

const pipeline = [
  { "$group": {
      "_id": "$customerId",
      "numPurchases": { "$sum": 1 },
      "numItemsPurchased": { "$sum": { "$size": "$items" } }
  } },
  { "$addFields": {
      "averageNumItemsPurchased": {
        "$divide": ["$numItemsPurchased", "$numPurchases"]
      }
  } }
]
return purchasesCollection.aggregate(pipeline).toArray()
  .then(customers => {
    console.log(`Successfully grouped purchases for ${customers.length} customers.`)
    for(const customer of customers) {
      console.log(`customer: ${customer._id}`)
      console.log(`num purchases: ${customer.numPurchases}`)
      console.log(`total items purchased: ${customer.numItemsPurchased}`)
      console.log(`average items per purchase: ${customer.averageNumItemsPurchased}`)
    }
    return customers
  })
  .catch(err => console.error(`Failed to group purchases by customer: ${err}`))

Find Data with Atlas Search

You can run Atlas Search queries on a collection with collection.aggregate() and the $search aggregation stage.

Important

App Services performs $search operations as a system user and enforces field-level rules on the returned search results. This means that a user may search on a field for which they do not have read access. In this case, the search is based on the specified field but no returned documents include the field.

exports = async function searchMoviesAboutBaseball() {
  // 1. Get a reference to the collection you want to search.
  const movies = context.services
    .get("mongodb-atlas")
    .db("sample_mflix")
    .collection("movies");
  // 2. Run an aggregation with $search as the first stage.
  const baseballMovies = await movies
    .aggregate([
      {
        $search: {
          text: {
            query: "baseball",
            path: "plot",
          },
        },
      },
      {
        $limit: 5,
      },
      {
        $project: {
          _id: 0,
          title: 1,
          plot: 1,
        },
      },
    ])
    .toArray();
  return baseballMovies;
};

{
  "plot" : "A trio of guys try and make up for missed
  opportunities in childhood by forming a three-player
  baseball team to compete against standard children
  baseball squads.",
  "title" : "The Benchwarmers"
}
{
  "plot" : "A young boy is bequeathed the ownership of a
  professional baseball team.",
  "title" : "Little Big League"
}
{
  "plot" : "A trained chimpanzee plays third base for a
  minor-league baseball team.",
  "title" : "Ed"
}
{
  "plot" : "The story of the life and career of the famed
  baseball player, Lou Gehrig.",
  "title" : "The Pride of the Yankees"
}
{
  "plot" : "Babe Ruth becomes a baseball legend but is
  unheroic to those who know him.",
  "title" : "The Babe"
}

Note

$$SEARCH_META Variable Availability

The $$SEARCH_META aggregation variable is only available for functions that run as system or if the first role on the searched collection has its apply_when and read expressions set to true.

If neither of these two scenarios apply, $$SEARCH_META is undefined and the aggregation will fail.

Aggregation Stages

Filter Documents

You can use the $match stage to filter incoming documents using standard MongoDB query syntax.

{
  "$match": {
    "<Field Name>": <Query Expression>,
    ...
  }
}

Example

The following $match stage filters incoming documents to include only those where the graduation_year field has a value between 2019 and 2024, inclusive.

{
  "$match": {
    "graduation_year": {
      "$gte": 2019,
      "$lte": 2024
    },
  }
}

Group Documents

You can use the $group stage to aggregate summary data for groups of one or more documents. MongoDB groups documents based on the _id expression.

Note

You can reference a specific document field by prefixing the field name with a $.

{
  "$group": {
    "_id": <Group By Expression>,
    "<Field Name>": <Aggregation Expression>,
    ...
  }
}

Example

The following $group stage groups documents by the value of their customerId field and calculates the number of purchase documents that each customerId appears in.

{
  "$group": {
    "_id": "$customerId",
    "numPurchases": { "$sum": 1 }
  }
}

Project Document Fields

You can use the $project stage to include or omit specific fields from documents or to calculate new fields using aggregation operators. To include a field, set its value to 1. To omit a field, set its value to 0.

Note

You cannot simultaneously omit and include fields other than _id. If you explicitly include a field other than _id, any fields you did not explicitly include are automatically omitted (and vice-versa).

{
  "$project": {
    "<Field Name>": <0 | 1 | Expression>,
    ...
  }
}

Example

The following $project stage omits the _id field, includes the customerId field, and creates a new field named numItems where the value is the number of documents in the items array:

{
  "$project": {
    "_id": 0,
    "customerId": 1,
    "numItems": { "$sum": { "$size": "$items" } }
  }
}

Add Fields to Documents

You can use the $addFields stage to add new fields with calculated values using aggregation operators.

Note

$addFields is similar to $project but does not allow you to include or omit fields.

Example

The following $addFields stages creates a new field named numItems where the value is the number of documents in the items array:

{
  "$addFields": {
    "numItems": { "$sum": { "$size": "$items" } }
  }
}

Unwind Array Values

You can use the $unwind stage to aggregate individual elements of array fields. When you unwind an array field, MongoDB copies each document once for each element of the array field but replaces the array value with the array element in each copy.

{
  $unwind: {
    path: <Array Field Path>,
    includeArrayIndex: <string>,
    preserveNullAndEmptyArrays: <boolean>
  }
}

Example

The following $unwind stage creates a new document for each element of the items array in each document. It also adds a field called itemIndex to each new document that specifies the element's position index in the original array:

{
  "$unwind": {
    "path": "$items",
    "includeArrayIndex": "itemIndex"
  }
}

Consider the following document from the purchases collection:

{
  _id: 123,
  customerId: 24601,
  items: [
    { name: "Baseball", quantity: 5 },
    { name: "Baseball Mitt", quantity: 1 },
    { name: "Baseball Bat", quantity: 1 },
  ]
}

If we apply the example $unwind stage to this document, the stage outputs the following three documents:

{
  _id: 123,
  customerId: 24601,
  itemIndex: 0,
  items: { name: "Baseball", quantity: 5 }
}, {
  _id: 123,
  customerId: 24601,
  itemIndex: 1,
  items: { name: "Baseball Mitt", quantity: 1 }
}, {
  _id: 123,
  customerId: 24601,
  itemIndex: 2,
  items: { name: "Baseball Bat", quantity: 1 }
}

Back

Write

MongoDB API Reference