Difference between revisions of "Perl one-liners"

Latest revision as of 10:39, 3 December 2008

A lot of nice things can be said about one-liners in Perl, but one example should suffice to give an understanding of their power and beauty:

One-liners

The following line reads the STDIN, concatenates a line at the end and prints to STDOUT. Its useful to concatenate arguments in STDOUT before calling a command:

>dir /B | perl -we"print<>;print'du'"

The following line reads the filename.txt content as the standard input and outputs the substituted content:

perl -p -e "s/old text/new text/gi;" < filename.txt

The next line reads the filename.txt content and writes the substituted content while keeping a backup of the original file (.bak):

perl -pi.bak -e "s/old text/new text/gi;" filename.txt

The next line reads an input file for a match. If match is found then splits the line by '/' char and prints the last element.

perl -F\/ -lane perl "if (m|MATCH|{$a=pop @F; print $a;})" < input.txt

The next lines reads input file and prints ok if a match is found:

perl -nwe"if(m/MATCH/gi){print'Ok';}" < input.txt

The next line reads stdinput, splits it, and prints it in uppercase:

> echo EV_prefix.tmp | perl -F[_\.] -wlane'pop@F;$a=pop@F;print uc($a);'
PREFIX

The next line extracts values from xml tags:

> curl http://www.cnn.com | perl -ne 'm/>([^<>]*?)<\// && print$1."\n"' | sed -e '/^$/d'

Utilities

Rename htm to html files

In windows:

>dir /B /S | perl -wlne"/([^ ]+)\.htm$/i&&rename$1.'.htm',$1.'.html'"

In linux:

>find | grep htm | perl -wlne'/([^ ]+)\.htm$/i&&rename$1.".htm",$1.".html"'

Copy files matching MATCH to SUBST

In windows:

>dir /B /S | perl -MFile::Copy -wlne"/(([^ ]+)MATCH([^ ]+))/&&copy$1,$2.'SUBST'.$3"

In linux:

>find | grep MATCH | perl -MFile::Copy -wlne'/(([^ ]+)MATCH([^ ]+))/&&copy$1,$2."SUBST".$3'

Parsing MOHID code for keywords

This little one-liner utility allows to parse MOHID code for keywords and default values:

> more ModuleGeometry.F90 | perl -wlne"m/keyword.+'(.+)'/gi && print $1; m/default.+=(.+),/gi && print $1;"

Lookahead and lookbehind

This little perl program illustrates the use of lookahead(?=) and lookbehind(?<=) (as shown in I'mMike):

$ perl -e '$num = 8927369280;'\
> '$num =~ s/(?<=\d)(?=(\d\d\d)+$)/,/g;'\
> 'print $num, "\n"'
8,927,369,280

Troubleshoots

Q:Why my perl one-liners work in windows and don't work in linux, or vice-versa?
A: Under windows you must encapsule the command in double-quotes (""), 
whereas in linux you must encapsule commands in single-quotes (' ').

See also

• Perl
• Regexp
• Bash