Import .csv into MongoDB creating arrays with objects - arrays

Question

I am starting to familiarize with MongoDB and now i have to import a relative large .csv file into a Mongo collection. So far I was able to import the .csv file without big problems (either with mongoimport or through Studio 3T). However, I would like to create some arrays for each specific document. Maybe by explaining with an example my data I could make it clearer.
worker_id | gender | birth | employer_id | start | end | type |
---------------------------------------------------------------------------
CAD1213 | M | 1990 | WRF1119 | 15jun2018 | 31dec2018 | 11.1.1 |
CAD1213 | M | 1990 | WAC1134 | 1jan2019 | 31dec2019 | 12.1.1 |
CAF1456 | F | 1972 | WAC1134 | 15aug1988 | 17sep2016 | 55.0.1 |
CAR5567 | F | 1991 | WER9280 | 15jun2018 | 31dec2018 | 1.1.1 |
CAR5578 | F | 1989 | WHI1176 | 15jun2011 | 1nov2011 | 12.2.5 |
CAR5578 | F | 1989 | WHI1176 | 2nov2011 | 31dec2012 | 12.2.5 |
This is just a mere example with few fields (columns). I have data for job contracts where: worker_id is the the identification of the worker, gender the sex, birth refers to the birth year of the worker, employer_id is the identification of each employer, start and end the beginning and the end of each specific work contract, respectively, while type is the specific work contract. The whole database obviously is characterize by more fields.
With the classic .csv import, for each line I would obtain a different document with specific _ids and each column would represent the fields. However, I would like to have as '_id' (although not necessarily) the value of the worker_id. As you can see, some worker_id are repeated in the column (in a file of over 17 millions of row you can imagine how many replica I could have), therefore it would be impossibile to select this field as _id.
Here it comes what I am struggling to obtain. I would like to create an array (for example called contracts) which contains information of each specific contract, hence objects, in this case the columns employer_id, start, end, and type. For some workers we would have just an element in the array (CAF1456 and CAR5567 in the example) while for others more than one (CAD1213 and CAR5578 in the example). Therefore, an hypothetical document would be like the following one:
_id: "CAD1213"
gender: "M"
birth: "1990"
contracts: Array
0: Object
employer_id: "WRF1119"
start: "15jun2018"
end: "31dec2018"
type: "11.1.1"
1: Object
employer_id: "WAC1134"
start: "1jan1988"
end: "31dec2019"
type: "12.1.1"
_id: "CAF1456"
gender: "F"
birth: "1972"
contracts: Array
0: Object
employer_id: "WAC1134"
start: "15aug1988"
end: "17sep2016"
type: "55.0.1"
_id: "CAR5567"
gender: "F"
birth: "1991"
contracts: Array
0: Object
employer_id: "WER9280"
start: "15jun2018"
end: "31dec2018"
type: "1.1.1"
_id: "CAR5578"
gender: "F"
birth: "1989"
contracts: Array
0: Object
employer_id: "WHI1176"
start: "15jun2011"
end: "1nov2011"
type: "12.2.5"
1: Object
employer_id: "WHI1176"
start: "2nov2011"
end: "31dec2012"
type: "12.2.5"
As far as I know this procedure cannot be performed directly with mongoimport, then it has to be performed after importing the whole .csv document (I guess). I hope someone could provide me some hints, suggestions or links where I can find some help in achieving this structure. Moreover, if other structures of the documents could be more suitable for my example, I am eager to know.
Thank you in advance.