How to do it in one line?

r has asked for the wisdom of the Perl Monks concerning the following question:

Hi out there,
who of you can solve the following problem?

Our goal is to read from a non-formatted text file and print it out with all occurences of exactly two same characters in consecutive order enclosed in parentheses. Condition is to do it in an one-line perl script, if possible using regex (and not split because that´s too easy).

If our input file would be:

sssshjfdhgriuu
xxddvvggggaaaiii
ougur9s oknddpp
hgorsagnnnnccc
nkjl(())uh

... the output should be:

sssshjfdhgri(uu)
(xx)(dd)(vv)ggggaaaiii
ougur9s( )okn(dd)(pp)
hgorsagnnnnccc
nkjl((()()))uh

One of the non-working (and more than one line) attempts:

while (<>) {
  s/(.)((?!\1).)\2(?!\2)/$1\($2$2\)$3/g;
  s/^(.)\1(?!\1)/\($1$1\)$2/;
  print;
}
[download]

Looking forward to your answers,
r

Comment on How to do it in one line? Download Code

Replies are listed 'Best First'.
Re: How to do it in one line? by Courage (Parson) on Jul 03, 2002 at 22:40 UTC
seems like homejob, because it's useless but tuitfull. Anyway (on Win32 shell): `perl -wne "s/(.)\1+(?!\1)/length($&)>2?qq[$&]:qq[($1$1)]/ge;print" fil +e.txt` [download] on Unix just `perl -wne 's/(.)\1+(?!\1)/length($&)>2?qq[$&]:qq[($1$1)]/ge;print' fil +e.txt` [download] Got same result on your input! Courage, the Cowardly Dog.	[reply] [d/l] [select]
Re2: How to do it in one line? by blakem (Monsignor) on Jul 04, 2002 at 06:49 UTC
Same basic idea, but I shaved a few chars off of it.... `# 1 2 3 4 #2345678901234567890123456789012345678901234 perl -pe's/(.)\1(\1*)(?!\1)/$2?$&:"($&)"/ge' file.txt` [download] -Blake	[reply] [d/l]
Re: Re2: How to do it in one line? by r (Initiate) on Jul 04, 2002 at 10:22 UTC
Thanks a lot, although it is not a homejob ;-) Question is if it´s possible using exclusively regex (no ternary operator). Explore the unknown, know the known. -r	[reply]
Re: How to do it in one line? by Abigail-II (Bishop) on Jul 04, 2002 at 11:03 UTC
Not the shortest possible answer, but without using `/e`: `perl -pe's/(\G\|(.)(?!\2))((.)\4)(?!\4)/$2($3)/g' file` [download] Abigail	[reply] [d/l]
Re: Re: How to do it in one line? by BrowserUk (Patriarch) on Jul 08, 2002 at 15:16 UTC
I don't suppose that someone would consider explaining how this works for me? I spent a couple of hours on trying to understand this in conjunction with perlman:perlre and I am frankly lost.	[reply]
Re: How to do it in one line? by Abigail-II (Bishop) on Jul 08, 2002 at 16:28 UTC
`s/(\G # Either were we finished the previous time # (or at the beginning the first time). \| # Or (.)(?!\2)) # A character ($2) not followed by itself ( # Capture ($3) (.)\4(?!\4) # A character ($4) followed exactly once by + itself. ) # End capture ($3) /$2($3)/xg # Replace with $2 (what preceeded) followed + by # "($3)" (the duplicate).` [download] It boils down to that if you have exactly two occurances of a character, something else should preceed those chararcters, or they should be at the beginning of the string (or were we left off the previous time). Abigail	[reply] [d/l]