I need some help with a unix script to mv image files to subdirectories based on the 1st character of the filename. See example below...
/images/main
1191.jpg
9999.jpg
A101.jpg
A102.jpg
B201.jpg
c333.jpg
...
Z999.jpg
I would like to move to the following:
/images/main/1/1191.jpg
/images/main/9/9999.jpg
/images/main/a/A101.jpg
/images/main/a/A102.jpg
/images/main/b/B201.jpg
/images/main/c/c333.jpg
...
/images/main/z/Z999.jpg
Any help would be greatly appreciated.
Thanks!
Another one:
#!/bin/sh
dir="/images/main"
for file in *.jpg
do
sd="$dir"/`echo "$file"|cut -c 1`
mkdir -p "$sd"
mv "$file" "$sd"
done
Regards
franklin52 script is very good but needs a tweak to change the directory to lowercase.
tr "[A-Z]" "a-z"
BTW. Don't use this syntax because it is upset by single character filenames!
tr [A-Z] [a-z]
If you have a very large number of files (or the filenames contain space characters) the script will need some work to generate the list of files in a pipe without decending the directory tree. Depends on which unix and which version of "find" you have.
Worked great. Thanks for your help!
I know he already has his answer . . . but I see a lot of answers on here
that are overly complicated. The ksh itself has much of the functionality
of cut, awk and sed. For example:
# This creates a variable that is only 1 character-wide, left-justified,upper-case
typeset -u -L1 first_letter
cd images/main
/bin/ls *.jpg |
while read file ; do
first_letter=$file
mkdir $first_letter
/bin/mv $file $first_letter
done
This script has the added feature of being a single process.
Therefore, if it mattered, which it probably doesn't nowadays,
it would be the fastest way performance-wise to do this.
Worked great the 1st time, but now when I try add more files to the directory and re-run it just hangs. Is there an issue because the directory ($dir) has already been created? If so, how do I modify to check first to see if $dir exists?
You can run it like so:
ksh -xvf script_file 2>&1 | more
And see exactly where it's getting hung up.
Other answer to your question:
if [ ! -d $dir ]; then
mkdir $dir
fi
Thanks quirkasaurus!
It appears the original script worked the 1st time, but when I ran it again it created another subdirectory below the previous one. I tried running several times before I looked at the output and now I have a huge mess ...
/images/main/a/a/a/a/a/A101.jpg
/images/main/b/b/b/b/B201.jpg
Depending on where I killed the program there are a varying number of sub-directories.
How can I get back to everything in the /images/main directory and start all over again with an alternative suggestion?
Thanks for your patience on this thread and your help is greatly appreciated.
First, put everything back to normal:
mkdir /tmp/my_images
find images/main -name \*.jpg -exec /bin/cp {} /tmp/my_images \; -print
Verify that all the images are there:
ls /tmp/my_images
Nuke everything in images/main:
cd images/main
\rm -rf *
Put all the images back:
cp /tmp/my_images/* .
Now.
Don't use the solutions that used the find command.
Instead, use mine, posted earlier. This one will avoid
any recursive directory building problem.
HTH
HTH - Thank you! OK - Recovered!
I am now using your script. How do I modify to handle "Arguement List Too Long" error because I have 30,000 images in the directory?
typeset -u -L1 first_letter
/bin/ls |
grep '.jpg$' |
while read file ; do
first_letter=$file
mkdir $first_letter
/bin/mv $file $first_letter
done
With much help from quirkasaurus this problem was resolved using BASH script below:
#!/bin/sh
cd "$1"
/bin/ls |
grep '.jpg$' |
while read file ; do
first_letter=`echo $file | cut -c 1 | tr '[A-Z]' '[a-z]'`
mkdir -p $first_letter
/bin/mv $file $first_letter
done