Shell Programming and Scripting

I have directory where new sub directories and files being created every few minutes. The directories are like abc_date, def_date, ghi_date. I am looking to keep the latest 2 unique directories and delete everything else.
Here is what I have so far

This gives me unique names excluding the date.
ls -lt | awk '{ print $9 }' | cut -d_ -f1 | sort | uniq


This will give me the files that I want to delete for the uniq word abc
ls -lt |grep abc | sort -r | tail +3

My question is, how do i pass the output of first command to the second one and delete what is needed? Any help is greatly appreciated.

Clarify your question, what is the format of the names of created subdirectories and wich directories should be delete? Give some examples.

Regards

How's this? You'll still need logic to decide what to remove.

ls -lt | awk '{ print $9 }' | cut -d_ -f1 | sort | uniq | while read prefix x
do
ls -t ${prefix}* | sort -r | tail +3
done

I have directories being created by some automated scripts like this
Broker_01022009.15.23
Broker_03042009.12.15
Broker_03062009.16.15
Dealer_02172009.14.20
Dealer_01222009.17.45
Dealer_02272009.16.40


What I want is to leave the last 2 and delete the rest directories and their contents. In this ex. what should be left is
Broker_03062009.16.15
Broker_03042009.12.15
Dealer_02272009.16.40
Dealer_02172009.14.20


Thanks in advance.

Be careful when writing this type of script.
If the directories are being created by multiple concurrent processes rather than consecutive single processes you will need to be sure that a the process has finished before deleting the directory.
BTW: The while construct proposed by aixylinux is sound and works when there are no directories to delete. IMHO it needs a minor tweak to extract the directory name rather than the files in the directory.
Note the use of ls - one rather than ls - ell because we just want the name.
For example:

Code:

for PREFIX in "Broker_" "Dealer_"
do
       ls -1d ${PREFIX}* 2>/dev/null | sort -r | tail +3 | while read DIR
       do
             if [ -d "${DIR}" ]
             then
                     echo "Old directory: ${DIR}"
             fi
       done
done

Try this:

Code:

ls -lt | awk '/^d/{split($8,f,"_");a[f[1]]++}a[f[1]] > 2{system("ls -l "$8)}'

If the command outputs the directories you want to remove, replace ls -l with rm -r:

Code:

ls -lt | awk '/^d/{split($8,f,"_");a[f[1]]++}a[f[1]] > 2{system("rm -r "$8)}'

Regards