Aggregation: group date in nested documents (nested object) - mongodb

Question

I've a mongo database with user information. If new userdata is added I do a duplicate check and in case of a duplicate entry, I do not create a new document, but instead update the existing one with a nested node (under tracking) adding the timestamp and some other informations.
{
"_id": "5e95dee277dcc55e9d18bf1a",
"email": "test#test.com",
"tracking": [
{
"domain": "mydomain",
"subdomain": "",
"ip": "59.214.120.68",
"timestamp": "2020-03-21 20:06:12",
"externalID": "82"
},
{
"domain": "mydomain",
"subdomain": "",
"ip": "99.214.130.33",
"timestamp": "2020-03-26 18:43:01",
"externalID": "483"
},
{
"domain": "mydomain",
"subdomain": "",
"ip": "19.214.131.22",
"timestamp": "2020-03-26 18:48:42",
"externalID": "485"
}
]
}
Now I'm trying to aggregate the documents and group/count them by date. Is there any option how I can do this with diffrent number of nodes under tracking for each document?

Answer 1

You can do something like this.
{
$unwind: {
path: "$tracking",
preserveNullAndEmptyArrays: true
}
}, {
$group: {
_id: "$tracking.timestamp",
count: {
$sum: 1
}
}
}

Answer 2

It would be much better if you could store datetimes as actual datetime type and not string representations. But assuming you cannot for now, to group by the date component without the time (which is what I believe you seek) you can use substr:
db.foo.aggregate([
{$unwind: "$tracking"}
,{$group: {"_id": {$substr: [ "$tracking.timestamp", 0, 10] } , n: {$sum:1} }}
]);