$unionWith (aggregation)

Duplicate Results

The combined results from the previous stage and the $unionWith stage can include duplicates.

For example, create a suppliers collection and a warehouses collection:

db.suppliers.insertMany([
  { _id: 1, supplier: "Aardvark and Sons", state: "Texas" },
  { _id: 2, supplier: "Bears Run Amok.", state: "Colorado"},
  { _id: 3, supplier: "Squid Mark Inc. ", state: "Rhode Island" },
])

db.warehouses.insertMany([
  { _id: 1, warehouse: "A", region: "West", state: "California" },
  { _id: 2, warehouse: "B", region: "Central", state: "Colorado"},
  { _id: 3, warehouse: "C", region: "East", state: "Florida" },
])

The following aggregation combines the state field projection results from the suppliers and warehouse collections.

db.suppliers.aggregate([
   { $project: { state: 1, _id: 0 } },
   { $unionWith: { coll: "warehouses", pipeline: [ { $project: { state: 1, _id: 0 } } ]} }
])

The result set contains duplicates:

{ "state" : "Texas" }
{ "state" : "Colorado" }
{ "state" : "Rhode Island" }
{ "state" : "California" }
{ "state" : "Colorado" }
{ "state" : "Florida" }

To remove the duplicates, you can include a $group stage to group by the state field:

db.suppliers.aggregate([
   { $project: { state: 1, _id: 0 } },
   { $unionWith: { coll: "warehouses", pipeline: [ { $project: { state: 1, _id: 0 } } ]} },
   { $group: { _id: "$state" } }
])

The result set no longer contains duplicates:

 { "_id" : "California" }
 { "_id" : "Texas" }
 { "_id" : "Florida" }
 { "_id" : "Colorado" }
 { "_id" : "Rhode Island" }

$unionWith a Sharded Collection

If the $unionWith stage is part of the $lookup pipeline, the $unionWith coll cannot be sharded. For example, in the following aggregation operation, the inventory_q1 collection cannot be sharded:

db.suppliers.aggregate([
   {
      $lookup: {
         from: "warehouses",
         let: { order_item: "$item", order_qty: "$ordered" },
         pipeline: [
            ...
            { $unionWith: { coll: "inventory_q1", pipeline: [ ... ] } },
            ...
         ],
         as: "stockdata"
      }
   }
])

Collation

If the db.collection.aggregate() includes a collation, that collation is used for the operation, ignoring any other collations.

If the db.collection.aggregate() does not include a collation, the db.collection.aggregate() method uses the collation for the top-level collection/view on which the db.collection.aggregate() is run:

• If the $unionWith coll is a collection, its collation is ignored.

• If the $unionWith coll is a view, then its collation must match that of the top-level collection/view. Otherwise, the operation errors.

Atlas Search Support

Starting in MongoDB 6.0, you can specify the Atlas Search $search or $searchMeta stage in the $unionWith pipeline to search collections on the Atlas cluster. The $search or the $searchMeta stage must be the first stage inside the $unionWith pipeline.

To see an example of $unionWith with $search, see the Atlas Search tutorial Run an Atlas Search $search Query Using $unionWith.

Restrictions

Create Sales Reports from the Union of Yearly Data Collections

The following examples use the $unionWith stage to combine data and return results from multiple collections. In these examples, each collection contains a year of sales data.

Report 1: All Sales by Year and Stores and Items

The following aggregation creates a yearly sales report that lists all sales by quarter and stores. The pipeline uses $unionWith to combine documents from all four collections:

db.sales_2017.aggregate( [
   { $set: { _id: "2017" } },
   { $unionWith: { coll: "sales_2018", pipeline: [ { $set: { _id: "2018" } } ] } },
   { $unionWith: { coll: "sales_2019", pipeline: [ { $set: { _id: "2019" } } ] } },
   { $unionWith: { coll: "sales_2020", pipeline: [ { $set: { _id: "2020" } } ] } },
   { $sort: { _id: 1, store: 1, item: 1 } }
] )

Specifically, the aggregation pipeline uses:

• A $set stage to update the _id field to contain the year.

• A sequence of $unionWith stages to combine all documents from the four collections, each also using the $set stage on its documents.

• A $sort stage to sort by the _id (the year), the store, and item.

Pipeline output:

{ "_id" : "2017", "store" : "General Store", "item" : "Chocolates", "quantity" : 150 }
{ "_id" : "2017", "store" : "General Store", "item" : "Cookies", "quantity" : 100 }
{ "_id" : "2017", "store" : "General Store", "item" : "Pie", "quantity" : 10 }
{ "_id" : "2017", "store" : "ShopMart", "item" : "Chocolates", "quantity" : 50 }
{ "_id" : "2017", "store" : "ShopMart", "item" : "Cookies", "quantity" : 120 }
{ "_id" : "2017", "store" : "ShopMart", "item" : "Pie", "quantity" : 5 }
{ "_id" : "2018", "store" : "General Store", "item" : "Cheese", "quantity" : 30 }
{ "_id" : "2018", "store" : "General Store", "item" : "Chocolates", "quantity" : 125 }
{ "_id" : "2018", "store" : "General Store", "item" : "Cookies", "quantity" : 200 }
{ "_id" : "2018", "store" : "General Store", "item" : "Pie", "quantity" : 30 }
{ "_id" : "2018", "store" : "ShopMart", "item" : "Cheese", "quantity" : 50 }
{ "_id" : "2018", "store" : "ShopMart", "item" : "Chocolates", "quantity" : 150 }
{ "_id" : "2018", "store" : "ShopMart", "item" : "Cookies", "quantity" : 100 }
{ "_id" : "2018", "store" : "ShopMart", "item" : "Nuts", "quantity" : 100 }
{ "_id" : "2018", "store" : "ShopMart", "item" : "Pie", "quantity" : 25 }
{ "_id" : "2019", "store" : "General Store", "item" : "Cheese", "quantity" : 50 }
{ "_id" : "2019", "store" : "General Store", "item" : "Chocolates", "quantity" : 125 }
{ "_id" : "2019", "store" : "General Store", "item" : "Cookies", "quantity" : 200 }
{ "_id" : "2019", "store" : "General Store", "item" : "Nuts", "quantity" : 80 }
{ "_id" : "2019", "store" : "General Store", "item" : "Pie", "quantity" : 50 }
{ "_id" : "2019", "store" : "ShopMart", "item" : "Cheese", "quantity" : 20 }
{ "_id" : "2019", "store" : "ShopMart", "item" : "Chocolates", "quantity" : 150 }
{ "_id" : "2019", "store" : "ShopMart", "item" : "Cookies", "quantity" : 100 }
{ "_id" : "2019", "store" : "ShopMart", "item" : "Nuts", "quantity" : 30 }
{ "_id" : "2019", "store" : "ShopMart", "item" : "Pie", "quantity" : 75 }
{ "_id" : "2020", "store" : "General Store", "item" : "Cheese", "quantity" : 100 }
{ "_id" : "2020", "store" : "General Store", "item" : "Chocolates", "quantity" : 200 }
{ "_id" : "2020", "store" : "General Store", "item" : "Cookies", "quantity" : 500 }
{ "_id" : "2020", "store" : "General Store", "item" : "Nuts", "quantity" : 100 }
{ "_id" : "2020", "store" : "General Store", "item" : "Pie", "quantity" : 100 }
{ "_id" : "2020", "store" : "ShopMart", "item" : "Cheese", "quantity" : 100 }
{ "_id" : "2020", "store" : "ShopMart", "item" : "Chocolates", "quantity" : 300 }
{ "_id" : "2020", "store" : "ShopMart", "item" : "Cookies", "quantity" : 400 }
{ "_id" : "2020", "store" : "ShopMart", "item" : "Nuts", "quantity" : 200 }
{ "_id" : "2020", "store" : "ShopMart", "item" : "Pie", "quantity" : 100 }

Report 2: Aggregated Sales by Items

The following aggregation creates a sales report that lists the sales quantity per item. The pipeline uses $unionWith to combine documents from all four years:

db.sales_2017.aggregate( [
   { $unionWith: "sales_2018" },
   { $unionWith: "sales_2019" },
   { $unionWith: "sales_2020" },
   { $group: { _id: "$item", total: { $sum: "$quantity" } } },
   { $sort: { total: -1 } }
] )

• The sequence of $unionWith stages retrieve documents from the specified collections into the pipeline:

• The $group stage groups by the item field and uses $sum to calculate the total sales quantity per item.

• The $sort stage orders the documents by descending total.

Pipeline output:

{ "_id" : "Cookies", "total" : 1720 }
{ "_id" : "Chocolates", "total" : 1250 }
{ "_id" : "Nuts", "total" : 510 }
{ "_id" : "Pie", "total" : 395 }
{ "_id" : "Cheese", "total" : 350 }