Find files older than X and Count them.

[mackman]

I have log files that should be parsed and then deleted by a script on a regular basis. Sometimes things don't work for a variety of reasons and the log files sit and sit and are never dealt with.

What I need is a small script that can give me the files older than X days and a count of those files.

What I have so far helps me take care of things manually but I need a little automation in my life

Here is what I have:


I can count all the files in the necessary directories recursively with this: ls -laR | wc -l

And I can find all the files that are older than 10 days that haven't been deleted yet by doing this: find /home/mike/logs -type f -mtime +10

But how do I put both of them into a script that will just give me the end number of both?

Can someone help me out here?

[clowenstein]

If you want file names and a count of them, use find(1) to get the names and gawk(1) to print and count.

$ find /home/mike/logs -type f -mtime +10 -print | gawk '{print $0}END{print NR}'

If you want more information per file, you can use the "-ls" operator of find(1).

$ find /home/mike/logs -type f -mtime +10 -ls | gawk '{print $0}END{print NR}'

This will give you even more information than "ls -l".

[mackman]

Ok, that was a big help! Now I have it working.

How could I put that in a tiny .sh script that would echo the result?

So the script would run and give me the result?

[scm]

Simplest way to count things is to use the program that's designed to do just that - wc

find /home/mike/logs -type f -mtime +10 -print | wc -l

And to print the result:

echo "Number of files: " $(find /home/mike/logs -type f -mtime +10 -print | wc -l)

Do them as separate arguments to get echo to lose the leading spaces that wc puts on its counts.

[clowenstein]

That's where the original poster started. But I believe he wanted file names and other information as well as the count.

[clowenstein]

If you want file names and a count of them, use find(1) to get the names and gawk(1) to print and count.

$ find /home/mike/logs -type f -mtime +10 -print | gawk '{print $0}END{print NR}'

If you want more information per file, you can use the "-ls" operator of find(1).

$ find /home/mike/logs -type f -mtime +10 -ls | gawk '{print $0}END{print NR}'

This will give you even more information than "ls -l".

Ok, that was a big help! Now I have it working.

How could I put that in a tiny .sh script that would echo the result?

So the script would run and give me the result?

Script:

# /bin/sh
find $1 -type f -mtime +10 -print \
| gawk '{print $0}END{print NR}'

Save as findscript.sh, put in a place that is in your $PATH. Ask if you don't understand $PATH.

$ chmod +x findscript.sh

Run as:

$ findscript.sh /home/mike/logs

If you want a bit of decoration on the results, change the gawk statement to:

gawk '{print $0}END{print "Number of files = " NR}'