With this function, I get the error
test: Wrong type argument: stringp, 2
Is it good to pass prefix this way? Thought that the argument prefix would be filled up with the prefix value.
(defun test (prefix)
"TODO"
(interactive "P")
(message prefix)
(message "Test"))
(defun foo (raw num)
(interactive "P\np")
(message "RAW: %S, Numeric: %S" raw num))
Now try M-x foo with different prefix args. The message will tell you how interactive handles P and p.
Your error is from function message, which C-h f will tell you expects a string as its first arg. Your first call to message calls it with whatever you used as the raw prefix argument.
M-x set-variable RET debug-on-error RET t enters the debugger when an error is raised, so you can see directly what your problem is.
Read about prefix arguments. In the Elisp manual (C-h i m elisp), i prefix argument takes you to node Prefix Command Arguments, which tells you all about it.
And if you're going to write Elisp code you owe it to yourself to read the manual An Introduction to Programming in Emacs Lisp, also available with C-h i.
raw and num for tho same prefix argument. For interpreting prefix argument, I find that the raw prefix argument is better to work with.
(cond ((equal raw '(4)) (statements for no C-u condition)?
'(4), then there was a C-u prefix. But why not elaborate your function in the light of your findings and try out the cond? Just use (message "No C-u") for the consequent e.g. It would be quicker to try out dozens of variations in the time that you are waiting for an answer here (which might never come).