UNIX for Dummies Questions & Answers

Hi all,

I'm having a problem with some basic piping issues...

I have been able to get in a directory and ls | gsed in order to list every N file for instance:

Code:

ls | gsed -n '2~5p'

The thing is I want to be able to copy the output files to a new directory. Basically directory /all has a bunch of jpegs in there. I need to copy every third one, starting at #2 to /selects.

Can some kind person point me on the way with this. I am on OS X but installing gsed as the OS X sed seems inconsistent with the examples I am using to get the initial output...

THANK YOU

OSX's default sed is an ancient BSD incarnation which doesn't even support things like \n. gsed is a good idea.

I don't know how "2~5p" is supposed to work though, trying it on /etc/passwd I actually get more lines out of it than I put in!

How about

Code:

ls | awk '(NR%3)==0' | while read LINE
do
        echo cp "$LINE" /path/to/dest/
done

One of the simplest ways to use awk is an expression which tells it whether to print the current line or not. NR is the number of records. So it will print the current line only when the line number divided by 3 has a remainder of 0 -- so line 3, 6, 9, 12, ...

Then you feed it into a while loop and move files one by one.

If you wanted to avoid running cp 30,000 times, you could pipe them into xargs/tar:

Code:

ls | awk '(NR%3)==0' | xargs -I {} tar -cf - | tar -C /path/to/dest/ -xf -

running 'tar -cf - file1 file2 file3' bundles files into a tape archive printed directly to standard output. xargs converts file names in a pipeline into arguments fed into tar. Piping that directly into tar -xf extracts the files. -C tells it where. So it pipes the contents of all the files between the two tars and creates them where you want.

Thank you so much Corona688 - that really helped me get moving this morning. Used your second example with perfect results.

Just beware that if there are too many files, tar will get run more than once, causing the pipe to break early, so some may not be copied. My first example is technically safer, if less efficient.

ok, good to know. this was an input of about 750 files, outputting 250 and from what I can tell at first glance seems to have worked well. I'll keep an eye on it. Thank you again for the help and lesson.