Yeah I think @Darren_Murphy tends to group by column type. I have been grouping my columns by function instead. For example, I have a table that lists students. I have a Name group that contains last name, first name, combined name, and other various ways that I format name. Then I have a Contact group for info, such as address, phone, email, etc. I have an Info Group that contain other miscellaneous info about the students. And, I have a GroupEmail group which handles all of the logic that I need for building and sending a group email to a select number of students that the coach chooses.
In another app, I have a series of columns that I needed to repeat 6 times as the function needs to be repeated. Each set performs a specific function as a combined group, and the result from that group is used for the next group. With grouping, I can easily find the correct column within the correct group.
There are a lot of ways to use grouping. Just depends on your preference and how you intend to organize your data.