Hello;

I have a directory that is collecting log and act files. I need to write a script that will remove these files once they are 30 days old. I have read through a number of threads on this site that have given me a great deal of information. However I have what seems to be a unique situation. I can't use the -atime or the -mtime commands because not all the files that haven't been modified or accessed can be deleted. There are non log or act files in this directory. One thing that may help is that the log and act files use yyyymmdd.log or yyyymmdd.act file naming system. IE 20050623.log or 20050623.act. How can I use this to generate my script? Thank you very much in advance for your knowledge and help.

Do this first to make sure it's getting what you want.


Then do this if it goes according to your criteria:

There are a fw problems with that find.

1. "*.act and *.log" will not match anything
2. It will start an rm process per file

Something like the following would be better:

Note I don't understand the OPs comment about mtime not working.
Note mtime is initialised with the date the file was created.

mtime is changed if the file is modified after the creation date ... you could have a file that has been there 5 years but modified twice daily so it's mtime will always be "-1" ...

IMUO --- the only difference in most cases between using "|xargs" as compared to "-exec" is the use of 2 commands instead of 1 ... in a directory with a small number of files, the use of either form is dictated more by the user's preference than anything else ...

xargs is a more general/scalable solution to many problems.

Having reread the OP, I think he meant "have" rather than "haven't" in the following:

"I can't use the -atime or the -mtime commands because not all the files that haven't been modified or accessed can be deleted."

So how about using -ctime ?
I know c means change not create, but that should
be the same.

Try this piece of code if you want to be 100% sure.

for filename in list
do
file_dt=`echo $filename | awk '{print substr($1,length($1)-12,8)}'`
today_dt=`date '+%Y%m%d'`
(( diff=$today_dt-$file_dt))

if [ $diff -gt 30 ]; then
do whatever u want
....
fi
done

using find and rm combination will be more fast.
but if you are not too sure, u can use this.

I have it working. I had to make a couple of minor tweaks but it works great now. I would like to thank everyone who replied. Everyone's responses where very clear and made this little task very easy. Again thank you all very very much.