I have a ASP.NET Web Api project, and have implemented repository and unitOfWork pattern.

I currently have a big sql query that I want to implement in code, it includes a lot of tables and joins.

My current solution is just fetching data from repositories one by one, returning an in-memory value or IEnumerable (not a IQueryable as it defeats the purpose of repositories) and just joining them later and projecting them to a DTO.

But this makes 20 round trips to the database for 20 tables im using, and its very inefficient.

How can I improve performance (mainly focused on lowering number of round trips to database) by following repository pattern?