MongoDB remove elements depending on element before (Iterating) - arrays

Question

I want to remove ($unset) elements from my MongoDB Objects with condition if the same Object has a similar element.
My Object:
{
"_id": "5eabf8b144345b36b00bfbaa",
"ranktime": [{
"pos":"2",
"datum":"Mon May 05 2020 12:22:52 GMT+0200 (GMT+02:00)",
"source":"SOURCE2"
},{
"pos":"1",
"datum":"Fri May 01 2020 12:23:10 GMT+0200 (GMT+02:00)",
"source":"SOURCE1"
},{
"pos":"37",
"datum":"Fri May 01 2020 12:25:14 GMT+0200 (GMT+02:00)",
"source":"SOURCE2"
},{
"pos":"12",
"datum":"Fri May 01 2020 12:25:14 GMT+0200 (GMT+02:00)",
"source":"SOURCE2"
},{
"pos":"37",
"datum":"Fri May 01 2020 18:45:27 GMT+0200 (GMT+02:00)",
"source":"SOURCE2"
}]
}
So I want to remove the entry in ranktime if ranktime.source == "SOURCE2" and if the date is the same as with the object before. Actually I have to iterate through the single elements of ranktime. Is this possible in MongoDB ?
The Expected outcome would be:
{
"_id": "5eabf8b144345b36b00bfbaa",
"ranktime": [{
"pos":"2",
"datum":"Mon May 05 2020 12:22:52 GMT+0200 (GMT+02:00)",
"source":"SOURCE2"
},{
"pos":"1",
"datum":"Fri May 01 2020 12:23:10 GMT+0200 (GMT+02:00)",
"source":"SOURCE1"
},{
"pos":"37",
"datum":"Fri May 01 2020 12:25:14 GMT+0200 (GMT+02:00)",
"source":"SOURCE2"
}]
}

Answer 1

Basically you can use $reduce to process an array and define previous element using $let and $arrayElemAt statements. The new $set syntax allows you to use aggregation within update statement:
db.col.updateMany({}, [
{
$set: {
ranktime: {
$reduce: {
input: "$ranktime",
initialValue: [],
in: {
$let: {
vars: { last: { $arrayElemAt: [ "$$value", -1 ] } },
in: {
$cond: [
{
$and: [
{ "$eq": [ "$$last.source", "SOURCE2" ] },
{ "$eq": [ { $substr: [ "$$last.datum", 0, 15 ] }, { $substr: [ "$$this.datum", 0, 15 ] } ] },
]
},
"$$value",
{ $concatArrays: [ "$$value", [ "$$this" ] ] }
]
}
}
}
}
}
}
}
])
Aggregation Example