I think Perl one-liners are still super useful. They are small Perl programs that are run directly from command line. Like this one from the Kubernetes job documentation:
perl -Mbignum=bpi -wle 'print bpi(2000)' # calculate PI to 2000 digitsperl is the Perl language interpreter. -M
and -wle are command line switches (or flags or options)
that modify the perl’s behaviour. See below for explanation
of what they mean. The string within the quotes is the Perl code that
gets executed. In this case it uses the bpi subroutine from
the bignum module to
calculate the PI with accuracy of 2000 digits. The command will take a
while to finish.
Switches
These are some of the most used command line switches:
-e '<code>'– execute<code>-E '<code>'– Execute<code>enabling new features for your version of Perl-w– enable warnings (generally advisable)-p– loop through lines, reading and printing them (in-script equivalent:while (<>) { [<code>] print })-n– loop through lines, reading but not printing them-l– print a newline ($/actually) after each line of output and chomp input newlines if used with-nor-p-i[<.ext>]– edit files in-place, optionally create backups with<.ext>extension-a– autosplit the$_default variable into@Farray (space is the default separator, change it with-F, ex.-F:)-M<module>[=<subroutine>,...]– load subroutine(s) from a Module
See perlrun for more.
Cut
Cut out 2nd and 1st space separated field (column):
$ cat birthdays.txt
2890-09-22 Bilbo Baggins
2968-09-22 Frodo Baggins
$ perl -wlane 'print join " ", @F[1,0]' birthdays.txt
Bilbo 2890-09-22
Frodo 2968-09-22The field numbering starts at 0. We use join to put a space between the cut out fields.
Search
Find lines in logs that contain error or warning:
perl -wne '/error|warning/i && print' /var/log/*.logThe thing between slashes is a regular expression. It
means match string error or string warning
anywhere in the log line. i says to Perl to ignore the
case. So it will match ERROR, error, Warning etc. If the regex finds a
match (i.e. evaluates to true) the && logical
operator runs the print statement that will print the line
containing the match.
Get IP addresses from logs:
journalctl --since "00:00" | perl -wlne '/((?:\d{1,3}\.){3}\d{1,3})/ && print $1' | \
sort | uniq > /tmp/ips.txtThe IP address regex explained:
( # capturing parenthesis to be retrieved via $1
(?: # non capturing parenthesis, only for grouping
\d{1,3} # one to three decimal numbers
\. # literal dot
){3} # three times all within innermost parenthesis
\d{1,3} # one to three decimal numbers
) For a more serious program where you want to cover possible edge cases you should use a well tested module Regexp::Common as suggested by PerlMonks.
Replace
Replace /bin/sh with /bin/bash and emit the
transformed passwd file to STDOUT:
perl -wpe 's#/bin/sh$#/bin/bash#' /etc/passwdWe used # instead of / as delimeter for
better readibility since the strings themselves contain slashes.
$ means end of the string.
Replace colour with color in all text
files. The original files will be kept with .bak
suffix:
perl -i.bak -wpe 's/colour/color/g' *.txtg (global) means replace all occurences in a string not
just the first one.
Convert between DOS and Unix newline:
perl -i -wpe 's/\r//' <file1> <file2> ... # dos-to-unix
perl -i -wpe 's/$/\r/' <file1> <file2> ... # unix-to-dosCalculate
Calculate the total size of log files older than 30 days:
find /opt/splunk/syslog/ -iname "*log*" -type f -mtime +30 | \
perl -wlne '$sum += (stat)[7]}{print $sum'The stat
function returns a 13-element list of status info about a file. We take
the 8th element (with index 7) which is the size of a file.
We loop over the found files and add the size of each into the
$sum variable. The handy Eskimo
Greeting Operator is for printing the $sum when the
loop is over (suggested by PerlMonks).
More
For a deeper dive see Famous Perl One-Liners Explained. If you want a book have a look at Minimal Perl for UNIX and Linux People.