The function i'm looking for has to return the index of the first , that is out of a pair of " ".
For example, with the sequence
{ " h i , a l l " : 3 , " h o w , i s " : " x " }
'( #\{ #\" #\h #\i #\, #\a #\l ... )
The function should return 11, not 4 (first occurrence of comma) because it is between " ".
I tried with this:
(defun control-comma (x p)
(cond ((eql (car x) #\")
(control-comma (subseq x (+ (position #\" x :start (+ 1 p)) 1)) p))
((eql (car x) #\,)
p)
(t
(control-comma (cdr x) (+ 1 p)))
)
)
Using x as list of input and p as a 0-parameter to count the position, but it doesn't work and seems to be far away from the solution i'm looking for.
Thank you for every suggestion.
Instead of defining a complex function, I suggest you to use the predefined position-if operator:
(defun first-comma (string start)
(let ((in-double-quote nil))
(position-if
(lambda (x)
(case x
((#\") (progn (setf in-double-quote (not in-double-quote)) nil))
((#\,) (not in-double-quote))))
string
:start start)))
CL-USER> (first-comma (coerce "{ \"hi, all\" : 3, \"how, is\" : \"x\" }" 'list) 0)
15
A more complex, recursive solution based again on the idea of scanning the input list one character at time, is given by the following function, where the state “inside double-quote” is encoded through a couple of recursive local functions:
(defun fist-comma (x pos)
(labels ((looking-for-comma (x pos)
(cond ((null x) nil)
((eql (car x) #\,) pos)
((eql (car x) #\") (looking-for-double-quote (cdr x) (1+ pos)))
(t (looking-for-comma (cdr x) (1+ pos)))))
(looking-for-double-quote (x pos)
(cond ((null x) nil)
((eql (car x) #\") (looking-for-comma (cdr x) (1+ pos)))
(t (looking-for-double-quote (cdr x) (1+ pos))))))
(looking-for-comma (nthcdr pos x) pos)))
Finally, note that in both the above functions one should take into account possible escaping of the double quote with appropriate means.