KSH script
 

Anyone has any idea how to do script this:

This will keep two directories whose first fields are the same.

I can do that by:

ls | awk -F '_' ' {print $1} '


If there are more than two dir for each field it will delete anything older than 60 days.

I can't figure out how to pick up two or more dir whose $1 look the same and not the rest.

Then I can rm -r exec the older ones by their mtime or atime or whatevertime, I have to look that up.

Any thoughts?

What do you mean by pick-up?
I think you need to clarify your requirements.

I agree feed a man a fish theory. That is why I only need help with this one section. By pick up I mean in a variable. How do I get a count of each individual file whose first fields are different. In other words I need to keep 2 files that are alike and rm the rest that are older by their date stamp. There are many many files with the same name. I only need the two newest one.

find / -type f -name "name_*" |xargs ls -tr | tail -n 2


That, or I still don't get what you wanted ...

Cheers,
Tink

Here's an example:

this is the $ls

en_05062010_1055 ResMorStorefrontAuthor_05062010_1109
en_05172010_0730 ResMorWebCSRSIT_05062010_1101
fr_05062010_1059 ResMorWebCSRSIT_05172010_0741
fr_05172010_0734 Tomcat_Ager.sh



ls | awk -F '_' ' {print $1} ' gives me this so I can sort them by whatever is in front of the _ ( underscore)

en
en
fr
fr
ResMorStorefront
ResMorStorefrontAuthor
ResMorWebCSRSIT
ResMorWebCSRSIT

I need to keep two of the files from each set of files and clip off anything older than the two newer ones.


Thanks in advance

It's funny when I understand the commands clearer than the question you're asking ;)

This is where I'm getting lost...

-C

Heh ...

Same ... another stab in the dark:
If you tack an "xargs rm" onto the line with the "head" it
should delete the oldest files.



Cheers,
Tink

Well I am going to go on a limb here and say that awk can probably do all of it if I am reading the file names correctly??

Apart from the Tomcat file the rest appear to have a date and time stamp after the initial name of the file so I am presuming
the ls will sort them logically (someone can shoot me down if I am wrong here), so I would use:

Just remove the pipe before rm to check if the right names will be removed.

That would work if the files had iso-dates in the names, in which
case date == ASCII sort order.

With MMDDYYYY (who dreamt that up, anyway!? ;D) you could have
older files with a higher ASCII sort order than the date value last.



Cheers,
Tink

thanks guys i appreciate the help

@Tinkster
While I am the first to say I have never really understood this date ordering, although the US sure likes it,
I was wondering if you could provide an example of where it would not work?
Not doubting you just couldn't think of an example ;)

Good enough?
No matter whether you sort ASCII or ASCII reverse, you
miss out on one of the 2010 dates.



Cheers,
Tink

If this is solved can you please mark it as such?



Cheers,
Tink