I'm making a script that will transfer some files from CD to harddrive. The files on the cd are in the format of:
I would like to end up with a file in the /tmp/ directory called _1.txt with these two entries:
This file will be used as a user selectable wildcard list.

Here's what I have so far:
What would be the correct syntax for awk or sed to match this: [0-9][0-9]???200[0-9] and just pipe it to "sort -u" so that it can be written to the /tmp/_1.txt file?

--Tony

if the file format is so regular, why not just use cut?

ls -1 /mnt/cdrom/[0-9][0-9]???200[0-9]* | cut -c 12-20 | sort | uniq > /tmp/_1.txt

otherwise...my sed is not so great, but something like this should strip the leading path

ls -1 /mnt/cdrom/[0-9][0-9]???200[0-9]* | sed 's/.*\///g'

but I'd still be inclined to use cut for the postfix

HTH

Thanks man!

Here's what I ended up doing:

-- Tony

ok. now that I thinl about it some more, I don't know why you need the -l flag to ls. nor do I get why you need the second grep.

oh, btw - be careful of uniq on it's own - always best to use sort | uniq (i.e. don't trust ls to sort properly for you)

cheers

I put the "1" flag (that's a number one ... not the letter L) just to make sure there is one entry per line. The grep is to make sure that what I get actually does start with a number ... during my testing I got one entry that was an empty space and another that was a strange character.

Thanks for your help

--Tony

The "-1" (one file per line option) is unnecessary, as when ls output is not a terminal, but redirected to a file or piped to another command, this is the default behaviour.

Also, I don't understand why you use 16-24 vs 12-20 which match for me the path you give.

Finally, the 'grep -e "^[0-9]"' still seems to me redundant, as the files already start by a number anyway. I would be curious to see how did you got an empty space or "strange characters", perhaps have you directories matching the ls pattern ?

ummm code or whatever:

#/bin/sh
#written in notepad with no way of testing this - user beware!

#ooooo... command line options!
#I suppose you could put the destination dir in as a command line option too...
sourceDir=$1
find $sourceDir -type f -iname '*.tar.gz' > /destination/dir/fileList.lst

eachList=`cat fileList.lst | tr "\n" " "`

for each in $eachList do ;
cp $each /destination/dir/ ;
echo `basename $each | cut -c 1-9` >> /destination/dir/baseList.lst ;
done

cat /destination/dir/baseList.lst | sort | uniq > /destination/dir/_1.txt

rm /destination/dir/fileList.lst /destination/dir/baseList.lst

echo -e "done! \n"

that might not be what you were looking for but there it is... as much for my own amusement as your use....