| Until a decent parser for arc becomes available (whether regexp-based or not) I've created the following regexp parser, which simply calls out to perl at the command line (tested on linux only) Here's the code: (def mmatch (regex fmt fname)
(tostring:system:string "perl -0ne 'print " (esc fmt) " while (/" regex "/igs)' " fname))
(def match (regex fname)
(mmatch regex "$1\n" fname))
That's all there is!These functions support the full perl regexp syntax :-) Example #1- Let's say you have a file with lots of junk and want to extract anything that looks like an email address. Use the "match" function, which finds a single subexpression multiple times: arc> (prn:match "([a-zA-Z0-9]+@[a-zA-Z0-9]+\\.[a-zA-Z0-9]+)" "~/temp/foo.txt")
bob@gmail.com
lauren@yahoo.com
Example #2- If you want to break each email down into multiple subexpressions at the same time, simply call the "mmatch" function. It requires an extra parameter that controls how we want the output to be formatted. In this example, we want the three parts of the email address broken out as items in a list: arc> (readall:mmatch "([a-zA-Z0-9]+)@([a-zA-Z0-9]+)\\.([a-zA-Z0-9]+)" "($1 $2 $3)\n" "~/temp/foo.txt")
((bob gmail com) (lauren yahoo com))
| 
