group files by strings in dir

Melsync · 10-25-2005, 04:45 PM

In a folder there are some files that have the first six characters of their names identical.
geis43*.txt
geis43*.xml
.
rayl18*.jpg
rayl18*.txt
rayl18*.xml
I need to apply some commands on each group of characters, like zipping the ones with a specific regex in their basenames, for instance, and dumping the others.
I've thought about declaring `ls .` as an array and declaring the basename of each file as another array. Then I'd loop an if test to compare each substring element of the arrays basename.
Not being any good in scripting, I suspect that there may be an easier way!

itsme86 · 10-25-2005, 04:52 PM

What's wrong with something like: zip foo.zip geis43*

?

Melsync · 10-25-2005, 05:08 PM

The problem is that the directory contains hundreds of files and I may change them from time to time, even daily.

Melsync · 12-16-2005, 12:40 PM

I suspect that the solution would be to chop the first 6 characters off declaring a variable with ${filename:0:5} and then a loop with chopped names:

Code:

do
       if [ -d ${i##*} ]
       then
           mv $i ${i##*}
       else
           mkdir ${i##*}
           mv $i ${i##*}
       fi
done
zip ${i##*}.zip ${i##*}*
rm `ls ${i##.*}* | grep -v ".zip"`

How do I define the variable containing the first 6 characters of the filename only?

Thanks.

Mel

sirclif · 12-16-2005, 01:52 PM

how about this...

Code:

#! /bin/bash

for base in `ls | sed "s/\(......\).*// | uniq`
do
   gzip $base* $base.zip
done

this will list all files in the working directory (if the script is in your $PATH variable, it will list all files in the directory you call it from), picks off the first 6 characters, then gets rid of any duplicate entries. so it will loop through all unique six character sequences and zip up all the files whose first six characters match into a file.

Melsync · 12-17-2005, 09:37 AM

Thanks, sirclif, but the code doesn't return anything.
I'm trying to correct the syntax.

Quote:

for base in .
do
togzip=`ls | sed -e 's/$......$.*//' $base | uniq`
gzip $tozip* $tozip.zip
done

Probably I've made a number of mistakes up there.

sirclif · 12-17-2005, 10:30 AM

oops, i left something out of the sed command. try this...

Code:

#! /bin/bash

for base in `ls | sed "s/\(......\).*/\1/ | uniq`
do
   gzip $base* $base.zip
done

my bad.

Melsync · 12-17-2005, 12:53 PM

Quote:

#! /bin/bash

for base in `ls | sed "s/$......$.*/\1/" | uniq`
do
zip $base.zip $base*
done

did the trick for me.
Thank you very, very much