• View
• Edit
• History
• Print

Linux

Linux/Unix snippets and tips:

Sed and Awk

Make

This diff's each file in two directories (as opposed to just diff'ing the directories ls - l creates a list from one dir, nawk then converts this to a two column file list n in xargs specifies the number of args to read at once

ls -l ../src/ | nawk '/^d/ { print $9 " ../src/" $9 }' | xargs -n 2 diff $1 $2

This is an alternative to the above tell xargs to use one line at once - seems nicer ls -l ../src/ | nawk '/^d/ { print $9 " ../src/" $9 }' | xargs -l diff $1 $2

build directory structure ls /some/dir | grep ".*/" | xargs mkdir

copy files from subdirs of parent dir to this dir (must have subdirs in place) find .. -name "*.F" | sed 's/..//' | xargs -I '{}' -t cp ../'{}' ./'{}'

tar selected files tar cvf ../shmuser/src.tar */{*.{F},*.{h},Make*.*}

Count files in a directory - use wc (word count) which reports lines, words and letters use -l to get line count

ls f* | wc -l

Create a tar based on find

find . -type f -name "*.java" | xargs tar rvf myfile.tar

Shell script to remove files of 0 length handy if a redirect yeilds no output

# Compare the file results
files="file1 file2"
for file in $files
do
#cmp does binary file compare
cmp dir1/${file} dir2/${file}  > $file.dif

if [ ! -s $file.dif ]
then
rm -f $file.dif
fi

if [ $? -ne 0 ]
then
echo "Differences found in" ${file}.dat
else
echo "No differences found in" ${file}.dat
fi
done

Awk/Shell script find files of given length

FS9=`ls -l fort.9 |awk '{print $5}'`
if [ $FS8 -eq 0 ] || [ $FS9 -gt 100 ]
then
...
fi

Get the absolute version of a relative path

ABS_DIR=`(cd ../../../..; pwd)`

Change case of filenames
for i in `ls`
do cp $i `echo $i | tr '[:upper:]' '[:lower:]'`;
done

Split a file into parts
csplit -k big-file '/regexp'/ '{100}'

Note the ' ' around the {} are NEEDED in CSH as CSH would otherwise interpret the {} before it got to csplit. A prefix and number of digits on the split files can also be provided

Produce a tree listing for a directory structure
If you are on a Unix system that doesn't have a tree list then try this// Credit to the origiantor listed in the header

#######################################################
#  A better tree, based on tree script from           #
#  http://www.centerkey.com/tree                      #
#                                                     #
#######################################################

echo
if [ "$1" != "" ]  #if parameter exists, use as base folder
then cd $1
fi
pwd
ls -R | grep ":$" |   \
#following may need to be all on one line
sed -e 's/:$//'  | awk -F/ \\
' { printf "|";  for (i=1; i<=NF-2; i++) { printf "   ", $i; } \\
        {if(NF>2)printf " +- ";else printf "--"} print $NF }'


if [ `ls -F -1 | grep "/" | wc -l` = 0 ]   # check if no folders
then echo "   -> no sub-directories"
fi
echo
exit

csh

If you are unfortunate to end up using csh (CShell?) then here are a few things to make life a bit easier - to set a shortened version of your path as the prompt

Create an awk file containing:
pwd | awk -F/ '{ if(NF < 4 )print ; else {for (i=NF-3; i<=NF; i++) printf "%s/", $i}; }'

and aliases for
alias setprompt 'set prompt="`whoami`@`uname -n`:`~/shortPathPrompt.awk`> "' alias cd 'chdir \!* && setprompt'

Turn on filename completion with
set filec
use Esc (not tab) to complete a file/dirctory or Ctrl+D to list available completions