Skip to main content

Aggregating data

QuickStart Module

This quickstart module gives an a comprehensive overview on aggregating data using the MongoDB aggregation pipeline.

Using ODSL to create aggregation pipelines​

An aggregation pipeline consists of an array of stages that are executed in order to produce an output array of resultant documents.

You can easily construct a MongoDB aggregation pipeline in the ODSL language using the aggregate command, here is a simple example:

aggregate ${object:"m101:sample_training.zips"}
    match state = "NY" and city = "ALBANY"
end

Aggregation Stages​

The following MongoDB Aggregation Stages are fully supported within the ODSL language:

All other pipeline stages can be used in ODSL in JSON format, e.g.

aggregate ${object}
    { "$collStats": { "storageStats": { } } }
end

Examples​

Below are a set of examples which show a variety of aggregations, aggregation stages and functions

Match and count​

aggregate ${object:"m101:sample_training.zips"}
    match state = "NY" and city = "ALBANY" and pop > 5000
    count "zips"
end

Using group and sort​

aggregate ${object:"m101:sample_training.companies"}
    group _id="$founded_year", num=sum(1)
    sort num asc
end

Using addFields and project​

aggregate ${object:"m101:sample_training.companies"}
    match ipo != null
    addFields symbol="$ipo.stock_symbol"
    project name, symbol
end

Using geoNear, sort and limit​

lands_end = Point([-5.712946243042564, 50.0692015134188])
aggregate ${object:"m101:sample_geospatial.shipwrecks"}
    geoNear near=lands_end, distanceField="distanceFromLandsEndKm", spherical=true, distanceMultiplier=0.000156786
    project name, distanceFromLandsEndKm
    sort distanceFromLandsEndKm asc
    limit 5
end

Using lookup​

aggregate ${object:"m101:sample_mflix.movies"}
    lookup from "comments" localField "_id" foreignField "movie_id" as "comments"
    match comments > size(0)
    limit 1
end

Using bucket​

aggregate ${object:"m101:sample_training.companies"}
    match founded_year > 1980 and number_of_employees != null
    bucket "$number_of_employees" boundaries [ 0, 20, 50, 100, 500, 1000, Infinity ] 
end

Using autoBucket​

aggregate ${object:"m101:sample_training.companies"}
    match founded_year > 1980 and number_of_employees != null
    bucketAuto "$number_of_employees" buckets 5
end

Using unwind​

aggregate ${object:"m101:sample_training.companies"}
    match number_of_employees>100000
    project name, city="$offices.city"
    unwind "$city"
    match city="Tokyo"
end