MongoDb combine $facet results into a single result set.

mongodb-facet-combine.js

db.getCollection('list').aggregate([
    {
        $facet: {
            "events":[{
                $match: {
                   'type': 'Event'
                }
            }],
            "tasks": [{
                $match: {
                   'type': 'Task'
                }
            }]
        }
    },
    {$project: {activity:{$setUnion:['$events','$tasks']}}},
    {$unwind: '$activity'},
    {$replaceRoot: { newRoot: "$activity" }}
]);

This was very useful, thank you!

How would you apply skip and limit on the combined results?

How would you apply skip and limit on the combined results?

The scenario is not contrived. Salesforce has an activity type from which they extend to create events and tasks. However, often times you want to display all activity on a particular record, sorted correctly by some criteria. If you don't do it in the query using facet and setUnion you end up having to do extra logic on your server or in your client.

The documentation states that $setUnion removes duplicates. How does that affect the final output if the output from both $facet stages contains fields with identical names?