1 minute read

One task that often recurs in programming is the need to parse a string representation a number(or several numbers) and convert it to its numeric value. Parsing integer value in Common Lisp is fairly straightforward process - we have the built-in function PARSE-INTEGER:

(parse-integer "100") ;; => 100
(parse-integer "100" :radix 2) ;; => 4

As you can see the function allows you to parse a string representation of a number in an arbitrary base system(the default is the decimal). With the keyword argument :radix you can specify a base in the interval 2-36. The function has a few other fancy capabilities as well - like the ability to process only a part of the string that has been passed to it and to ignore junk in the input string. For all the gory details refer to the Lisp HyperSpec.

The problem that most Lisper face soon after is that there is no matching function PARSE-FLOAT or PARSE-DOUBLE. I’m not sure what technical reason is hidden beneath this design decision, but I know of simple way to parse floating point numbers non-the-less. It’s built around the READ(The R in REPL) function that allows you read any S-expression from a string form. The READ function then returns a Lisp object corresponding to the S-expression read. With that knowledge and the fact that READ accepts as an optional argument an input stream from which to read that S-expression(the default is the the standard input) we can write the following bit of parsing code:

(with-input-from-string (in "3.14")
  (read in))

Here we created an input stream that’s bound to the string “3.14” and read one S-expression from it - the floating point object 3.14.

We can even build a more general solution that parses several numbers in a string, regardless of their actual type(integer or floating point):

(with-input-from-string (in "3.14 5.646 4 9.6")
  (loop for x = (read in nil nil) while x collect x))
;; => (3.14 5.646 4 9.6)

Hopefully this short article has been helpful. You’ve also witnessed one of the practical benefits of having the code in Lisp represented as data.