Hi everyone

I'm quite new to Linux but still need to get the job done. My email server is running out of disk space.I need to delete some emails direct from the server and my first attempt is to delete emails dedicated to allstaff@mycompanyemail

I've grep the absalute path of those emails and store them into another files like this

grep –r allstaff@mycompanyemail* > /tmp/allstaff

Can anybody give the idea on how can i delete those emails by using the information in /tmp/allstaff. Should you have any other way is also can.

Thanks

grep –r allstaff@mycompanyemail* | xargs rm

find . -name 'allstaff@mycompanyemail*' | xargs rm

Never mind, misread post...

That command will not display anything. How if I want to display what file being deleted and the total size of deleted files? I may want to confirm that it will not delete the wrong file

Thanks

If you have them in /tmp/allstaff, presumably you have vetted them there. Simple variation on the above

Unfortunately, the thing is not as easy as I thought. My grep command was not only return the path but plus other info at the end of it like this

a/b/filename: for <allstaff@mycompanydomain>; Tue, 6 Mar 2007 12:05:41 +0800


I think I may need to remove the unnecessary character before can execute rm command. I'm not good in doing script. Please help

Thanks for your time

____________________________________________________________

I create some files in my own directory and do a test on it first. Hope that could consider BACKUP :>

To select a/b/filename part from this, use the command
"cut -d':' -f1"
The pipe sequence then becomes: cat | cut | xargs where the cat may be replaced with the grep or find commands.

To make sure that you don't delete anything by accident, you could add the -i option (interactive mode) to rm, which makes it ask for confirmation before actually deleting anything.
You could also use a harmless command, like echo or ls -l instead of the rm at the end of your pipe sequence.

An all-in-one solution could look something like this:

Thank you timmeke. It almost solved my problem. However there are few file names with double dot and numbers or characters like this :

a/d/filenamewith':'and'number'or'char':for <allstaff@mycompanydomain>; Tue, 6 Mar 2007 12:05:41 +0800

eg : a/d/fgfg:2,S:To: <allstaff@mycompanydomain>

Actually I don't really know how to use the option for cut command

Thanks

Maybe you can split up the strings on spaces rather than the colons ("double dots")?
eg use cut -d' ' -f1

-d => delimiter character
-f => cuts out certain fields (that are separated by the delimiter)

Note that "cut" only works on single character delimiters. So if we need to split up your lines on let's say the " To: " string, we'll need to look at "sed" or "awk", rather than "cut".

Ok. Thanks everyone for your time and assistence. I'm done!