$replaceWith (aggregation)
Definition
New in version 4.2.
Replaces the input document with the specified document. Theoperation replaces all existing fields in the input document,including the _id
field. With $replaceWith
, you canpromote an embedded document to the top-level. You can also specifya new document as the replacement.
The $replaceWith
is an alias for$replaceRoot
.
The $replaceWith
stage has the following form:
- { $replaceWith: <replacementDocument> }
The replacement document can be any valid expression that resolves to a document.For more information on expressions, seeExpressions.
Behavior
If the <replacementDocument>
is not a document,$replaceWith
errors and fails.
If the <replacementDocument>
resolves to a missing document (i.e.the document does not exist), $replaceWith
errors andfails. For example, create a collection with the followingdocuments:
- db.collection.insertMany([
- { "_id": 1, "name" : { "first" : "John", "last" : "Backus" } },
- { "_id": 2, "name" : { "first" : "John", "last" : "McCarthy" } },
- { "_id": 3, "name": { "first" : "Grace", "last" : "Hopper" } },
- { "_id": 4, "firstname": "Ole-Johan", "lastname" : "Dahl" },
- ])
Then the following $replaceWith
operation fails because oneof the document does not have the name
field:
- db.collection.aggregate([
- { $replaceWith: "$name" }
- ])
To avoid the error, you can use $mergeObjects
to mergethe name
document with some default document; for example:
- db.collection.aggregate([
- { $replaceWith: { $mergeObjects: [ { _id: "$_id", first: "", last: "" }, "$name" ] } }
- ])
Alternatively, you can skip the documents that are missing the name
field byincluding a $match
stage to check for existence of thedocument field before passing documents to the $replaceWith
stage:
- db.collection.aggregate([
- { $match: { name : { $exists: true, $not: { $type: "array" }, $type: "object" } } },
- { $replaceWith: "$name" }
- ])
Or, you can use $ifNull
expression to specify some otherdocument to be root; for example:
- db.collection.aggregate([
- { $replaceWith: { $ifNull: [ "$name", { _id: "$_id", missingName: true} ] } }
- ])
Examples
$replaceWith an Embedded Document Field
Create a collection named people
with the following documents:
- db.people.insertMany([
- { "_id" : 1, "name" : "Arlene", "age" : 34, "pets" : { "dogs" : 2, "cats" : 1 } },
- { "_id" : 2, "name" : "Sam", "age" : 41, "pets" : { "cats" : 1, "fish" : 3 } },
- { "_id" : 3, "name" : "Maria", "age" : 25 }
- ])
The following operation uses the $replaceWith
stage toreplace each input document with the result of a$mergeObjects
operation. The $mergeObjects
expression merges the specified default document with the pets
document.
- db.people.aggregate( [
- { $replaceWith: { $mergeObjects: [ { dogs: 0, cats: 0, birds: 0, fish: 0 }, "$pets" ] } }
- ] )
The operation returns the following results:
- { "dogs" : 2, "cats" : 1, "birds" : 0, "fish" : 0 }
- { "dogs" : 0, "cats" : 1, "birds" : 0, "fish" : 3 }
- { "dogs" : 0, "cats" : 0, "birds" : 0, "fish" : 0 }
$replaceWith a Document Nested in an Array
A collection named students
contains the following documents:
- db.students.insertMany([
- {
- "_id" : 1,
- "grades" : [
- { "test": 1, "grade" : 80, "mean" : 75, "std" : 6 },
- { "test": 2, "grade" : 85, "mean" : 90, "std" : 4 },
- { "test": 3, "grade" : 95, "mean" : 85, "std" : 6 }
- ]
- },
- {
- "_id" : 2,
- "grades" : [
- { "test": 1, "grade" : 90, "mean" : 75, "std" : 6 },
- { "test": 2, "grade" : 87, "mean" : 90, "std" : 3 },
- { "test": 3, "grade" : 91, "mean" : 85, "std" : 4 }
- ]
- }
- ])
The following operation promotes the embedded document(s) with thegrade
field greater than or equal to 90
to the top level:
- db.students.aggregate( [
- { $unwind: "$grades" },
- { $match: { "grades.grade" : { $gte: 90 } } },
- { $replaceWith: "$grades" }
- ] )
The operation returns the following results:
- { "test" : 3, "grade" : 95, "mean" : 85, "std" : 6 }
- { "test" : 1, "grade" : 90, "mean" : 75, "std" : 6 }
- { "test" : 3, "grade" : 91, "mean" : 85, "std" : 4 }
$replaceWith a Newly Created Document
Example 1
An example collection sales
is populated with the followingdocuments:
- db.sales.insertMany([
- { "_id" : 1, "item" : "butter", "price" : 10, "quantity": 2, date: ISODate("2019-03-01T08:00:00Z"), status: "C" },
- { "_id" : 2, "item" : "cream", "price" : 20, "quantity": 1, date: ISODate("2019-03-01T09:00:00Z"), status: "A" },
- { "_id" : 3, "item" : "jam", "price" : 5, "quantity": 10, date: ISODate("2019-03-15T09:00:00Z"), status: "C" },
- { "_id" : 4, "item" : "muffins", "price" : 5, "quantity": 10, date: ISODate("2019-03-15T09:00:00Z"), status: "C" }
- ])
Assume that for reporting purposes, you want to calculate for eachcompleted sale, the total amount as of the current report run time. Thefollowing operation finds all the sales with status C
and createsnew documents using the $replaceWith
stage. The$replaceWith
calculates the total amount as well as usesthe variable NOW
to get the current time.
- db.sales.aggregate([
- { $match: { status: "C" } },
- { $replaceWith: { _id: "$_id", item: "$item", amount: { $multiply: [ "$price", "$quantity"]}, status: "Complete", asofDate: "$$NOW" } }
- ])
The operation returns the following documents:
- { "_id" : 1, "item" : "butter", "amount" : 20, "status" : "Complete", "asofDate" : ISODate("2019-06-03T22:47:54.812Z") }
- { "_id" : 3, "item" : "jam", "amount" : 50, "status" : "Complete", "asofDate" : ISODate("2019-06-03T22:47:54.812Z") }
- { "_id" : 4, "item" : "muffins", "amount" : 50, "status" : "Complete", "asofDate" : ISODate("2019-06-03T22:47:54.812Z") }
Example 2
An example collection reportedsales
is populated with thereported sales information by quarter and regions:
- db.reportedsales.insertMany( [
- { _id: 1, quarter: "2019Q1", region: "A", qty: 400 },
- { _id: 2, quarter: "2019Q1", region: "B", qty: 550 },
- { _id: 3, quarter: "2019Q1", region: "C", qty: 1000 },
- { _id: 4, quarter: "2019Q2", region: "A", qty: 660 },
- { _id: 5, quarter: "2019Q2", region: "B", qty: 500 },
- { _id: 6, quarter: "2019Q2", region: "C", qty: 1200 }
- ] )
Assume that for reporting purposes, you want to view the reported salesdata by quarter; e.g.
- { "_id" : "2019Q1", "A" : 400, "B" : 550, "C" : 1000 }
To view the data grouped by quarter, you can use the followingaggregation pipeline:
- db.reportedsales.aggregate( [
- { $addFields: { obj: { k: "$region", v: "$qty" } } },
- { $group: { _id: "$quarter", items: { $push: "$obj" } } },
- { $project: { items2: { $concatArrays: [ [ { "k": "_id", "v": "$_id" } ], "$items" ] } } },
- { $replaceWith: { $arrayToObject: "$items2" } }
- ] )
- First stage:
- The
$addFields
stage adds a newobj
documentfield that defines the keyk
as the region value and thevaluev
as the quantity for that region. For example:
- { "_id" : 1, "quarter" : "2019Q1", "region" : "A", "qty" : 400, "obj" : { "k" : "A", "v" : 400 } }
- Second stage:
- The
$group
stage groups by the quarter and uses$push
to accumulate theobj
fields into a newitems
array field. For example:
- { "_id" : "2019Q1", "items" : [ { "k" : "A", "v" : 400 }, { "k" : "B", "v" : 550 }, { "k" : "C", "v" : 1000 } ] }
- Third stage:
- The
$project
stage uses$concatArrays
tocreate a new arrayitems2
that includes the_id
info and theelements from theitems
array:
- { "_id" : "2019Q1", "items2" : [ { "k" : "_id", "v" : "2019Q1" }, { "k" : "A", "v" : 400 }, { "k" : "B", "v" : 550 }, { "k" : "C", "v" : 1000 } ] }
- Fourth stage:
- The
$replaceWith
uses the$arrayToObject
to convert theitems2
into adocument, using the specified keyk
and valuev
pairs andoutputs that document to the next stage. For example:
- { "_id" : "2019Q1", "A" : 400, "B" : 550, "C" : 1000 }
The aggregation returns the following document:
- { "_id" : "2019Q1", "A" : 400, "B" : 550, "C" : 1000 }
- { "_id" : "2019Q2", "A" : 660, "B" : 500, "C" : 1200 }
$replaceWith a New Document Created from $$ROOT and a Default Document
Create a collection named contacts
with the following documents:
- db.contacts.insert([
- { "_id" : 1, name: "Fred", email: "fred@example.net" },
- { "_id" : 2, name: "Frank N. Stine", cell: "012-345-9999" },
- { "_id" : 3, name: "Gren Dell", cell: "987-654-3210", email: "beo@example.net" }
- ]);
The following operation uses $replaceWith
with$mergeObjects
to output current documents with defaultvalues for missing fields:
- db.contacts.aggregate([
- { $replaceWith: { $mergeObjects: [ { _id: "", name: "", email: "", cell: "", home: "" }, "$$ROOT" ] } }
- ])
The aggregation returns the following documents:
- { "_id" : 1, "name" : "Fred", "email" : "fred@example.net", "cell" : "", "home" : "" }
- { "_id" : 2, "name" : "Frank N. Stine", "email" : "", "cell" : "012-345-9999", "home" : "" }
- { "_id" : 3, "name" : "Gren Dell", "email" : "beo@example.net", "cell" : "", "home" : "987-654-3210" }