In Lisp style languages, a list is usually defined like this:

(list 1 2 3)

For the purposes of this challenge, all lists will only contain positive integers or other lists. We will also leave out the list keyword at the start, so the list will now look like this:

(1 2 3)

We can get the first element of a list by using car. For example:

(car (1 2 3))
==> 1

And we can get the original list with the first element removed with cdr:

(cdr (1 2 3))
==> (2 3)

Important: cdr will always return a list, even if that list would have a single element:

(cdr (1 2))
==> (2)
(car (cdr (1 2)))
==> 2

Lists can also be inside other lists:

(cdr (1 2 3 (4 5 6)))
==> (2 3 (4 5 6))

Write a program that returns code that uses car and cdr to return a certain integer in a list. In the code that your program returns, you can assume that the list is stored in l, the target integer is in l somewhere, and that all the integers are unique.


Input: (6 1 3) 3

Output: (car (cdr (cdr l)))

Input: (4 5 (1 2 (7) 9 (10 8 14))) 8

Output: (car (cdr (car (cdr (cdr (cdr (cdr (car (cdr (cdr l))))))))))

Input: (1 12 1992) 1

Output: (car l)

  • \$\begingroup\$ Can we take input with the integer first and the list second? \$\endgroup\$ Commented Sep 28, 2015 at 11:31
  • \$\begingroup\$ @MartinBüttner Sure. \$\endgroup\$
    – absinthe
    Commented Sep 28, 2015 at 11:32
  • \$\begingroup\$ What about (1 2 3) 16, shall we return ()? \$\endgroup\$
    – coredump
    Commented Sep 28, 2015 at 12:26
  • \$\begingroup\$ @coredump Good question. You can assume that the target integer will always be in the expression, so a case like (1 2 3) 16 will never show up. \$\endgroup\$
    – absinthe
    Commented Sep 28, 2015 at 12:27
  • \$\begingroup\$ Can we receive two inputs, one for the list and one for the integer? \$\endgroup\$
    – Blackhole
    Commented Sep 28, 2015 at 13:13

8 Answers 8


Common Lisp, 99

The following 99 bytes solution is a CL version of the nice Scheme answer.

(defun g(l n &optional(o'l))(if(eql n l)o(and(consp l)(or(g(car l)n`(car,o))(g(cdr l)n`(cdr,o))))))

I originally tried to make use of position and position-if, but it turned out to be not as compact as I would have liked (209 bytes):

(lambda(L x &aux(p'l))(labels((f(S &aux e)(cons(or(position x S)(position-if(lambda(y)(if(consp y)(setf e(f y))))S)(return-from f()))e)))(dolist(o(print(f L))p)(dotimes(i o)(setf p`(cdr,p)))(setf p`(car,p)))))


  (l x &aux (p 'l))
  (labels ((f (s &aux e)
              (or (position x s)
                   (lambda (y)
                     (if (consp y)
                         (setf e (f y))))
                  (return-from f nil))
    (dolist (o (print (f l)) p)
      (dotimes (i o) (setf p `(cdr ,p)))
      (setf p `(car ,p)))))


(funcall *fun* '(4 5 (1 2 (7) 9 (10 8 14))) 14)

The list is quoted, but if you really want, I can use a macro. The returned value is[1]:


For tests, I used to generate a lambda form where l was a variable:

(LAMBDA (#:G854) (CAR (CDR (CDR (CAR (CDR (CDR (CDR (CDR (CAR (CDR (CDR #:G854))))))))))))

Calling this with the original list returns 14.

[1] (caddar (cddddr (caddr l))) would be nice too

  • 2
    \$\begingroup\$ You answered a question about Lisp with Lisp! It's Lisp-ception! \$\endgroup\$
    – DanTheMan
    Commented Sep 29, 2015 at 1:02
  • 4
    \$\begingroup\$ @DanTheMan Lisp-ception is pretty much what defines Lisp ;-) \$\endgroup\$
    – coredump
    Commented Sep 29, 2015 at 5:27

Retina, 170 142 125 115 114 87 84 83 75 73 70 69 68 67 bytes

Yay, less than 50% of more than 100 bytes off my first attempt. :)

\b(.+)\b.* \1$
+`a *\)|\d

(c$2r $1)

To run the code from a single file, use the -s flag.

I'm still not convinced this is optimal... I won't have a lot of time over the next few days, I will add an explanation eventually.


Pyth, 62 bytes

JvXz"() ,][")u?qJQG&=J?K}Quu+GHNY<J1)hJtJ++XWK"(cdr "\d\aG\)\l

Try it online: Demonstration or Test Suite


The first bit JvXz"() ,][") replaces the chars "() " with the chars "[]," in the input string, which ends up in a representation of a Python-style list. I evaluate it and store it in J.

Then I reduce the string G = "l" with u...\l. I apply the inner function ... repeatedly to G, until the value of G doesn't change anymore and then print G.

The inner function does the following: If J is already equal to the input number, than don't modify G (?qJQG). Otherwise I'll flatten the the list J[:1] and check if the input number is in that list and save this to the variable K (K}Quu+GHNY<J1)). Notice that Pyth doesn't have a flatten operator, so this is takes quite a few bytes. If K is true, than I update J with J[0], otherwise with J[1:] (=J?KhJtJ). And then I replace G with "(cdr G)" and replace the d the a, if K is true (++XWK"(cdr "\d\aG\)).


Scheme (R5RS), 102 bytes

(let g((l(read))(n(read))(o'l))(if(pair? l)(or(g(car l)n`(car,o))(g(cdr l)n`(cdr,o)))(and(eq? n l)o)))

CJam, 59

q"()""[]"er~{:AL>{0jA1<e_-_A(?j'l/"(car l)"@{2'dt}&*}"l"?}j

Try it online


q                 read the input
"()""[]"er        replace parentheses with square brackets
~                 evaluate the string, pushing an array and a number
{…}j              calculate with memoized recursion using the array as the argument
                   and the number as the memozied value for argument 0
  :A              store the argument in A
  L>              practically, check if A is an array
                   if A is a (non-empty) array, compare with an empty array
                   (result 1, true)
                   if A is a number, slice the empty array from that position
                   (result [], false)
    {…}           if A is an array
      0j          get the memoized value for 0 (the number to search)
      A1<         slice A keeping only its first element
      e_          flatten array
      -           set difference - true iff the number was not in the array
      _           duplicate the result (this is the car/cdr indicator)
      A(          uncons A from left, resulting in the "cdr" followed by the "car"
      ?           choose the cdr if the number was not in the flattened first item,
                   else choose the car
      j           call the block recursively with the chosen value as the argument
      'l/         split the result around the 'l' character
      "(car l)"   push this string
      @           bring up the car/cdr indicator
      {…}&        if true (indicating cdr)
        2'dt      set the character in position 2 to 'd'
      *           join the split pieces using the resulting string as a separator
    "l"           else (if A is not an array) just push "l"
                   (we know that when we get to a number, it is the right number)
    ?             end if

PHP - 177 bytes

I've added some newlines for readability:

function f($a,$o,$n){foreach($a as$v){if($n===$v||$s=f($v,$o,$n))return
'(car '.($s?:$o).')';$o="(cdr $o)";}}function l($s,$n){echo f(eval(strtr
("return$s;",'() ','[],')),l,$n);}

Here is the ungolfed version:

function extractPhp($list, $output, $number)
    foreach ($list as $value)
        if (is_int($value))
            if ($value === $number) {
                return '(car '. $output .')';
            $subOutput = extractPhp($value, $output, $number);
            if ($subOutput !== null) {
                return '(car '. $subOutput .')';

        $output = '(cdr '. $output .')';

function extractLisp($stringList, $number)
    $phpCode = 'return '. strtr($stringList, '() ','[],') .';';
    $list = eval($phpCode);
    echo extractPhp($list, 'l', $number);

Haskell, 190 188 bytes

l "(4 5 (1 2 (7) 9 (10 8 14)))" 8

evaluates to

"(car (cdr (car (cdr (cdr (cdr (cdr (car (cdr (cdr l))))))))))"

l(h:s)n=c$i(show n)s""""
i n(h:s)t l|h>'/'&&h<':'=i n s(t++[h])l|t==n='a':l|h=='('=j$'a':l|h==')'=j$tail$dropWhile(=='d')l|0<1=j$'d':l where j=i n s""
c(h:s)="(c"++h:"r "++c s++")"
  • 1
    \$\begingroup\$ You can turn (and c in function c into a string: c(h:s)="(c"++h:... \$\endgroup\$
    – nimi
    Commented Sep 29, 2015 at 21:51
  • \$\begingroup\$ Wow, didn't think that would work with h being a Char! \$\endgroup\$ Commented Sep 30, 2015 at 9:07

Common Lisp, 168 155 bytes

Some stupid recursion thing, it could probably be condensed a bit more:

(lambda(l e)(labels((r(l o)(setf a(car l)d(cdr l)x`(car,o)y`(cdr,o))(if(equal e a)x(if(atom a)(r d y)(if(find e l)(r d y)(if d(r d y)(r a x)))))))(r l'l)))

Pretty printed:

(lambda (l e)
  (labels ((r (l o)
             (setf a (car l) d (cdr l)
                   x `(car ,o) y `(cdr ,o))
             (if (equal e a) x
                 (if (atom a)
                     (r d y)
                     (if (find e l)
                         (r d y)
                         (if d
                             (r d y)
                             (r a x)))))))
    (r l 'l)))

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.