"When I call PDF(2) I just see the 2nd integer numer from the last 1000x1. I would think I would see the 2nd 1000x1"
Your code does not use any indexing when allocating to the variable PDF, instead on each loop iteration your code simply replaces PDF completely, so in the end you will have only the vector(?) of data from the last loop iteration. That is the vector that you are indexing into.
Note how your other loops use indexing to allocate to the preallocated output array, e.g.:
data3(x)= ...
% ^^^ indexing to allocate into data3
whereas there is no indexing for allocating the results of the last loop:
PDF = test/length(i);
% ^^ where is the indexing?
Note that the loop iterator i will be a scalar and length of a scalar is one, so your code always divides by one.
Note that numbering variables like that is a sign that you are doing something wrong.
Note that copy-and-pasting code like that is a sign that you are doing something wrong.
Both of these conspire to make this task more verbose and more complex. Rather than writing separate code for importing the data from each file you should use a simple loop, just like the MATLAB documentation recommends:
Note how the examples use a loop and store all of the imported data in one cell array. You can then use indexing to access the file data (i.e. given the mydata cell array, the 2nd file data would be simply mydata{2}, exactly as you describe).
If the content of that cell array have suitable sizes then they could be concatenated into one matrix, e.g.:
mat = [mydata{:}]
