ok here is what i want to do

i run this command

ls -l |grep "^d"

it gives the complete listing of the directory

what i want is that some how i should be able to extract the names of the directories only that are output at the end of the ls -l command

same goes for the files

im not sure i understan what you mean by extract... but you could put them in a file

ls -l >>somefile

Where did you want to extract them to?

by extrating i mean

that when i get the output from the ls -l command i should be able to extract only the file names and not the rest of the rubbish which comes out of it.

bash-3.00$ ls -l
total 2881
drwx------ 2 jnsahibz mc00 512 Mar 14 16:15 CV
drwxr-xr-x 2 jnsahibz mc00 512 Mar 20 22:27 Desktop
drwx------ 3 jnsahibz mc00 512 Feb 17 11:28 Favorites
drwx------ 3 jnsahibz mc00 512 Mar 5 18:14 My Documents
drwx------ 3 jnsahibz mc00 512 Mar 22 14:49 assignment
drwxr-xr-x 3 jnsahibz mc00 512 Mar 22 13:18 bin
-rw------- 1 jnsahibz mc00 2931431 Mar 19 17:21 junaid.zip
drwx------ 2 jnsahibz mc00 512 Mar 19 19:59 mail
drwxr-xr-x 4 jnsahibz mc00 512 Mar 21 19:41 public_html
drwx------ 3 jnsahibz mc00 512 Mar 13 18:45 zakir naik


now u see at the end of every line there is the name of the directory or file

that is what i need.

the reason i am using ls -l is because i wan to separate files and directories. so i can do that using "d" or "-" to make out whether it is a file or a directory.

papa ur code works for extracting directory names. how about extracting file names?
papa plz explain how to do the same for extracing the names of the files

this is what i am trying to do but unable to extract the file name

files=`ls -l |grep "^-" |cut -f20,25 -d" "`

try find instead of ls

for directories:
for files:

This will work for your directories

ls -l | tr -s " " | grep "^d" | cut -d " " -f 9

-Prasanta

In my personal opinion (which is much like a butt, everyone has one), piping ls -l through grep, for this case, is kinda like buying a full nice suit just to get the tie.

To get files/dirs separately:
The beauty of find is that it provides a full path, relative to the search path. If that's not your bag, and all you need is the filename, then you can strip it using basename:
If any of the files or dirs will have spaces in the name, you'll need to adjust the value of IFS, to remove the space as a field separator. Here is another thread with information about using IFS

This script might help you, what i do here is read the contents of an entire directory, and since i'll be working with the sub-dirs I read them into an array and from there I can open each dir ( array element ) and do all sorts of stuff. The code that I used is being used for backup purposes, where some of the structure is more all less the same and where there are differences in the tree structure I'd make some exceptions and work around it.


x=`find -maxdepth 1 -type d | sed -e '1d' sed 's/.\///g' 1>$DIR_LIST`

declare -a dirs
dirs=(`cat "$DIR_LIST"`)

cnt=0
numElem=${#dirs[*]}

while [ $cnt -lt $numElem ]
do
if [ ${dirs[$cnt]} == "clients" ]; then
echo "Working on Clients ... "

if [ -e $TARGET_IBOUND"/clients" ];then
cd ${dirs[$cnt]}
clients
cd .. #closes clients
else
mkdir -p $TARGET_IBOUND/${dirs[$cnt]}
cd ${dirs[$cnt]}
clients
cd ..
fi
fi

## Do whatever you want here

cnt=`expr $cnt + 1`

done

thanks to all of those who took time to reply to me

the problem is that our teacher has clearly instructed us to keep away from the FIND command.

any one using this command will get a zero