Best practice to run only one sharepoint.files

Hi Guys,

I want to have one datasource from SharePoint.Files (Csv tables) and based on this build seperated queries. 

What is the best practice ? 

I should build CSV tables in a source like add column here with Csv.Document ?:

 
and from query just filter specific row and expand table column?

Or i should create a table from binary for each seperated query? 
I want to avoid refreshing in the loop SharePoint source a lot of times...

Best,
Jacek