Originally posted by: SystemAdmin

Any of u have command/script to delete files based on date,without using mtime

I used this command
#find ./* -prune -mtime +2 -print | xargs rm

Originally posted by: alethad

I have a for statement to calculate the number of days old to remove files. You should be able to modify it to fit your needs. I send output to a temp file for my own use. I also got an awk script(awkc) to help me with grep'ing columns which I use for alot of scripts so that I don't re-invent the wheel everytime I need to write a script.
<hr />
for i in $TARGET_DIR/$file.*
do
#echo $i
fd=`ls -al $i |awkc 7` #get file date

(( n = $dom - $fd )) #calculate days old
#echo $n

if (( $n >= 1 )); then #if greater than 1 day remove it
#echo $dom $fd $n
echo Removing $i >>/tmp/archive_remove.out
rm $i 2>>/tmp/archive_remove.out
#echo Removing $i
fi
done
<hr />
Hope that helps you.

Originally posted by: SystemAdmin

Can you please include the awkc script? THX.

Originally posted by: esv

what are you trying to accomplish? does the solution has to be written in ksh/awk combo? could perl be used??

it would be a lot easier to begin with your goal and constrains, explaining carefully what have you done, then we might be able to help much better.

best regards,
esv.

Originally posted by: alethad

Here's a copy of the awkc script file. I don't know perl very well but I collect different perl and unix scripts for my own use. I'm sure you could probably come up with something better if you know perl.

The default field separator is a space. But you can change it to fit your needs.
<hr />
#!/bin/ksh

1. awkc - print out one or more columns

p=\$$( echo $1 | sed 's/,/,\$/g' )
shift
eval "awk '{ print $p }'" $*

1. eof

<hr />
Examples of how to use the awkc. Just pipe any unix command to it to get whichever column of data you need.

ps -ef | grep $process |awkc 2

fd=`ls -al $i |awkc 7` #get file date
Hope that helps.

Originally posted by: alethad

For some reason my reply for the awkc script inserted a number where a pound sign should be.
So ignore the 1. in the middle of the script and put in a pound sign.

Sorry for the trouble. Not sure why it did that.

Originally posted by: kappa

Sorry, alethad,
I have missed the point:
Where is the difference between awkc 2 and awk '{print $2}' ?
You could write awk \'{print $1}\' # $1 holding the Positional Param: awkc 2 => $1=2
Is it the same? (awkc code: I have no idea what it could do.)

I prefer things like:
ls *.log | perl -lne 'unlink if -M > 30' # remove files older than 30 days

It is faster, because unlink (=rm) is a syscall , not a process.
But I see, you prefer awkc as a filter utility.

Originally posted by: esv

I cheat big time.

I create a flag file with the touch command and a specific date (use -t flag)

touch -t 200807010000 /tmp/file

then find and delete all files older than it.
find /tmp -type f \! -newer /tmp/file | xargs rm

good luck.
esv.

Originally posted by: kappa

#find ./* -prune -name date* -print | xargs rm
-> files based on date
-> without using mtime

Bad answer for a bad question.

Please, be precise!
What means "based on date"?
Why not mtime? (Do you need ctime, do you need find at all?)

Do you pay dollars for each word in your question?