Hello. Newbie is too advanced a description for me when it comes to Unix script writing, so I do everything the hard way by using multiple 'single' line commands.

I am trying to copy off log files to an 'archive' folder, then compress them to preserve disk space. However, I want to add a cron job that would do this nightly for any newly created log files.

I most familiar with the 'Find' command. I want to use that to copy files created just yesterday, nothing older. Then I figure I will use another cron job to compress the newly added files.


I've seen many suggestions on this site but they either contain a specific date value ex: "Apr 14" or the command includes everything that is yesterday's date plus older.

Here is my starting command:
find ./ -type f -ctime +7 -exec cp -p {} ./archive/ \;

I've manned 'find', 'ctime' and 'mtime' and I can't find what I'm looking for.

Does anyone have an answer? Or know if this is possible? If not possible, can someone please offer a script solution?

Thanks!!

You want mtime - time the file was last modified.
I use touch and -newer in find for exact results:

Hi Jim,

I tried your example - I created two files called dummy1 and dummy2, ran the touch command so they would have today & yesterday's date. Then tried your command listed below. I got a 'missing conjunction' error. What did I type incorrectly? Thanks for all your help. L

ucasd84:iclac:/home/iclac# ll dummy*
-rw-r--r-- 1 iclac lawson 1 May 27 00:00 dummy1
-rw-r--r-- 1 iclac lawson 43 May 28 13:28 dummy2

ucasd84:iclac:/home/iclac# find . \( -newer dummy1 -a ! -newer dummy2 \) ./t.awk ./filename ./dummy2
find: 0652-009 There is a missing conjunction

This just lists the files that fit the time constraint

Thank Jim I got that to work. Very cool.

However, in your example you are using the find command w/a file names (dummy1/dummy2) and date values that are known. In my problem directory I won't know the file names to use for the comparison. The application that creates the files, which contains a bug hence the reason for this script, uses assending # order for the file names. It can create anywhere from 1 to 1000+ files per day.
EX file names:
4120.wuerr dated 05/27/09
4121.wuerr dated 05/28/09
4122.wuerr dated 05/28/09

My brain isn't smart enough today to figure out how to use your very cool find command in this scenario. Any thoughts on this? Hopefully my question made sense.

Thanks, Leslie