Virtual Brain/Linux/text operations 1 sort listing numerically {code:none} ls -1 1.txt 10.txt 2.txt {code} but {code:none} ls -1 | sort -n 1.txt 2.txt 10.txt {code} :-) 1 column moving Will take file and print the (whitespace separated? didn't check this) first two columns of each line in reversed order (first column 2 then column 1). {code:none} awk '{ print $2 $1 }' file {code} 1 column selection Print only second to third column of comma separated text in file. {code:none} cut -f 2-3 -d , file {code} 1 replace text in multiple files Perl offers the easiest way to do in-place-editing of text files: {code:none} perl -pi -e 's/tobereplaced/withthis/' filename # for the cautious people: make backups before editing the file: perl -pi'.bak' -e 's/tobereplaced/withthis/' filename # therefore multiple files: find . -name filename -exec perl -pi -e 's/pattern/bar/' {} \; # same with sed (in place edit, extended regex) sed -i -r "s/tobereplaced/withthis/" filename {code} It's possible to use other characters than / to delimit the patterns, e.g. pipe (|) or semicolon. This helps if the pattern itself contains slashes. Note that pattern is a regexp. Thanks also to Stefan Tramm for this hints :-) http://hacks.oreilly.com/pub/h/73 Replacing several things in a file with the above shorthand can be time consuming as a new perl process is started for every file and every replacement. An elegant way is to use a script that contains all the replacement commands - for sure this works with perl, but I usually use sed for this: {code:none} # sed script 'deprecation-off.sed' could look like this: s/org.eclipse.jdt.core.compiler.problem.deprecation=warning/org.eclipse.jdt.core.compiler.problem.deprecation=ignore/ s/org.eclipse.jdt.core.compiler.problem.invalidJavadocTagsDeprecatedRef=enabled/org.eclipse.jdt.core.compiler.problem.invalidJavadocTagsDeprecatedRef=disabled/ # and a call to edit all eclipse settings files could look like this: find ./*/.settings -name org.eclipse.jdt.core.prefs -print -exec sed -i -f deprecation-off.sed {} \; {code} 1 bash string operations Just a few but really useful: {code:none} # length of a string ${#string} # some examples > echo ${USER} stefan > echo ${#USER} 6 > echo ${USER:2} efan > echo ${USER:2:1} e {code} Source: http://tldp.org/LDP/abs/html/ 2011-03-29 12:30:51.757 2005-01-12 21:28:01.54