Do I need to create a data model for arbitrary selections from the database (application architecture)?

Question

C++ / C#

I created a data model (classes), for each class I created tables in SQLite to store data. Further data entry created the view models and views. Tried to follow MVVM. Here everything seems to be clear and working.
Now it was required to create a certain report containing the data from five connected tables.
There were questions:
1. Is it necessary to create a class for this sample from the point of view of OOP and MVVM?
2. Where to place this class, in data models or view models?
3. There can be two or ten times more such random samples compared to the data storage model, how are these tasks architecturally implemented?
I work with the database without EF in this project.