Xref: utzoo alt.religion.computers:2267 comp.lang.perl:3424 Path: utzoo!utgpu!cs.utexas.edu!uunet!convex!dragonne.convex.com!news From: tchrist@convex.COM (Tom Christiansen) Newsgroups: alt.religion.computers,comp.lang.perl Subject: Re: wc clone Message-ID: <1990Dec20.193625.662@convex.com> Date: 20 Dec 90 19:36:25 GMT References: <1990Dec20.003046.7902@convex.com> <1990Dec20.003805.8017@convex.com> <1990Dec20.020632.10077@convex.com> Reply-To: tchrist@convex.COM (Tom Christiansen) Followup-To: comp.lang.perl Organization: CONVEX Software Development, Richardson, TX Lines: 55 Nntp-Posting-Host: pixel From the keyboard of bzs@world.std.com (Barry Shein): :Imagine, in perl, if you could insert any expression midway into a :pattern so whenever whitespace was hit you could increment $words :right there, let's say "/(\S+)@$words++@/" was a pattern which :incremented $words every time a space run was found, that's a common :thing in snobol. Your example is darn close, tho, at least the loop :has been eliminated via the use of /g, that was the spirit of the :thing. : :Just various ways to find mapcar nirvana... (Barry, are you sure you're not just leading me on? :-) Permit me to introduce you to an eval in another guise, the /e modifier: [this one even works -- the last one had a bug. :-(] 0 #!/usr/bin/perl -n 1 $chars += length; 2 s/\S+/$words++/eg; 3 next unless eof; 4 printf "%8d %8d %8d %s\n", $., $words, $chars, ($ARGV eq '-'?'':$ARGV); 5 $tlines += $.; $twords += words; $tchars += $chars; reset 'wc'; $. = 0; 6 next unless $files++ && eof(); 7 printf "%8d %8d %8d %s\n", $tlines, $twords, $tchars, "total"; While evals are pretty neat, this one does slow things down a lot. Putting it in the line loop like that makes this program run 3.5 times longer than with line two as simply: $words += s/\S+//g; Better to count when all done in this case, but there are lots of more complex and interesting things you can do with /e easily. Let's say you also want to create a count of all the distinct words and then print them out in descending numeric order by count at the end of each file: #!/usr/bin/perl -n s/\S+/$saw{$&}++/eg; next unless eof; sub down { $saw{$b} <=> $saw{$a}; } for (sort down keys %saw) { printf "%8d %s\n", $saw{$_}, $_; }; Is that the kind of thing you are looking for, Barry? We've strayed pretty from religion here (I think:-), so I'm redirecting followups back into comp.lang.perl. --tom -- Tom Christiansen tchrist@convex.com convex!tchrist "With a kernel dive, all things are possible, but it sure makes it hard to look at yourself in the mirror the next morning." -me