Ad Hoc - trying to understand: missing numbers / numbers don't add up in 50k row reports

I'm trying to understand the "logic" of how reports are created in order to understand why we having incomplete reports.

I need to run large reports (50k rows by 100 columns), where the actual data set is far bigger than 50k x 100.

Let's say I want doc ids (50k rows) by page views of 100 countries.

Imagine column 1 is doc ids, colum 2 is US page views, column 3 is UK page views etc.

If I then run the same report ONLY for the UK, I discover high-hitting UK docs that are not in the big report.  

Given this, I assume the logic for building the report is:

1 Build the list of the top 50k docs for the country with the highest total page views (eg US) 

2 Display list of top 50k doc ids for US in column 1

3 Display page views of each doc id for the US in column 2

4 Column 3 (UK): for each doc id in column 1, display page views for the UK.

So, I assume that if a very high-hitting doc only exists in the UK but the US has the highest total page views, the high-hitting UK doc won't appear in the list.

If my assumption is right, is there a way to use total page views of all doc ids worldwide as the criteria for building column 1?

nick