Count duplicate dates occurrences - mongodb

Question

I've a collection in MongoDB of objects with this structure:
{
"_id": "ID",
"email": "EMAIL",
"name": "Foo",
"surname": "Bar",
"orders": [
{
"createdAt": "2019-09-09T07:30:25.575Z"
},
{
"createdAt": "2019-10-30T14:20:04.849Z"
},
{
"createdAt": "2019-10-30T16:38:27.271Z"
},
{
"createdAt": "2020-01-03T15:49:39.614Z"
},
],
}
I need to count all duplicates "createdAt" and distinct it with changing date format.
The result should be like below:
{
"_id": "ID",
"email": "EMAIL",
"name": "Foo",
"surname": "Bar",
"orders": [
{
"date": "2019-09-09",
"total": 1,
},
{
"date": "2019-10-30",
"total": 2,
},
{
"date": "2020-01-03",
"total": 1,
},
],
}
I tried with $unwind orders.createdAt in db.collection.aggregate() but i've no idea how can i get this result.
Thanks in advance.

Answer 1

Try this on for size. Given this data:
db.foo.insert([
{
"_id": "ID",
"email": "EMAIL", "name": "Foo", "surname": "Bar",
"orders": [
{ "createdAt": new Date("2019-09-09T07:30:25.575Z") },
{ "createdAt": new Date("2019-10-30T14:20:04.849Z") },
{ "createdAt": new Date("2019-10-30T16:38:27.271Z") },
{ "createdAt": new Date("2020-01-03T15:49:39.614Z") }
]
},
{
"_id": "ID2",
"email": "EMAIL2", "name": "Bin", "surname": "Baz",
"orders": [
{ "createdAt": new Date("2019-09-09T07:30:25.575Z") },
{ "createdAt": new Date("2020-10-30T14:20:04.849Z") },
{ "createdAt": new Date("2020-10-30T16:38:27.271Z") },
{ "createdAt": new Date("2020-10-30T15:49:39.614Z") }
]
}
]);
This agg:
db.foo.aggregate([
{$unwind: "$orders"}
// First $group is on just the Y-M-D part of the date plus the id.
// This will produce the basic info the OP seeks -- but not in the desired
// data structure:
,{$group: {
_id: {orig_id: "$_id", d: {$dateToString: {date: "$orders.createdAt", format: "%Y-%m-%d"}} },
n:{$sum:1} ,
email: {$first: "$email"},
name: {$first: "$name"},
surname: {$first: "$surname"}
}}
// The group is not guaranteed to preserve the order of the dates. So now that
// the basic agg is done, reorder by DATE. _id.d is a Y-M-D string but fortunately
// that sorts correctly for our purposes:
,{$sort: {"_id.d":1}}
// ...so in the second $group, we pluck just the id from the id+YMD_date key and
// take the YMD_date+n and *push* it onto a new orders array to arrive at the
// desired data structure. We are not guaranteed the order of orig_id (e.g.
// ID or ID2) but for each id, the push *will* happen in the order of arrival -- which was
// sorted correctly in the prior stage! As an experiment, try changing the
// sort to -1 (reverse) and see what happens.
,{$group: {_id: "$_id.orig_id",
email: {$first: "$email"},
name: {$first: "$name"},
surname: {$first: "$surname"},
orders: {$push: {date: "$_id.d", total: "$n"}} }}
]);
yields this output:
{
"_id" : "ID",
"email" : "EMAIL",
"name" : "Foo",
"surname" : "Bar",
"orders" : [
{
"date" : "2019-09-09",
"total" : 1
},
{
"date" : "2019-10-30",
"total" : 2
},
{
"date" : "2020-01-03",
"total" : 1
}
]
}
{
"_id" : "ID2",
"email" : "EMAIL2",
"name" : "Bin",
"surname" : "Baz",
"orders" : [
{
"date" : "2019-09-09",
"total" : 1
},
{
"date" : "2020-10-30",
"total" : 3
}
]
}
If you are willing to have a slightly more complex return structure and some dupe data in return for greater dynamic behavior by not having to enumerate each field (e.g. field: {$first: "$field"} then you can do this:
db.foo.aggregate([
{$unwind: "$orders"}
,{$group: {
_id: {orig_id: "$_id", d: {$dateToString: {date: "$orders.createdAt", format: "%Y-%m-%d"}} },
n:{$sum:1} ,
ALL: {$first: "$$CURRENT"}
}}
,{$group: {_id: "$_id.orig_id",
ALL: {$first: "$ALL"},
orders: {$push: {date: "$_id.d", total: "$n"}} }}
]);
to yield this:
{
"_id" : "ID2",
"ALL" : {
"_id" : "ID2",
"email" : "EMAIL2",
"name" : "Bin",
"surname" : "Baz",
"orders" : {
"createdAt" : ISODate("2019-09-09T07:30:25.575Z")
}
},
"orders" : [
{
"date" : "2019-09-09",
"total" : 1
},
{
"date" : "2020-10-30",
"total" : 3
}
]
}
{
"_id" : "ID",
"ALL" : {
"_id" : "ID",
"email" : "EMAIL",
"name" : "Foo",
"surname" : "Bar",
"orders" : {
"createdAt" : ISODate("2019-10-30T14:20:04.849Z")
}
},
"orders" : [
{
"date" : "2019-10-30",
"total" : 2
},
{
"date" : "2020-01-03",
"total" : 1
},
{
"date" : "2019-09-09",
"total" : 1
}
]
}