That is very useful but not quite what I wanted.
In your output, the primary sort are the numbers and the secondary are the letters.
I wanted it the other way around. Therefore all the data with A or a at the front would come before the data with B and so on. Therefore it would be (to start)