File handling script

For a first script, this is great, and I am impressed.

• your structure is neat
• you are using loops properly
• you are using variable substitution which is 'advanced'

The script appears that it will work for all the standard cases, and, I have seen much worse in 'production' code.

So, as a beginner, well done.

What's next?

• the edge cases are going to kill this script - spaces in file names.
• the logging and error handling could be better
• I recommend basename and dirname instead of the variable regexes you use:
  • dirname: http://linux.die.net/man/1/dirname
  • basename: http://linux.die.net/man/1/basename (note that basename path/to/file.pdf .pdf returns file)
• consistency on for->do loop - you have do on the next line once, and then for ... ; do the next time. Either is fine, but be consistent.

All of the error-handling, logging, and the spaces-in-files can be solved easily by using a function for running commands. Also, the "$@" construct in bash is 'special'

Consider the bash function:

function runcmd {
    echo "$@"
    # Fail if the command fails...
    "$@" || exit 1
}

So, this function will print the command, and then will run it, and exit the script if the command fails.

Now, instead of structures like:

            echo cp $working_directory/$folder_name/$f $new_filename
            cp $working_directory/$folder_name/$f $new_filename

you can do:

            runcmd cp "$working_directory/$folder_name/$f" $new_filename

and you will still get the echo, you will get the new error handling, and you will handle spaces in names just fine.

As @rolfl said, the script will fail if the directories contain spaces. In any case, a for loop over the output of the find command is really not pretty. You can get around the issue by changing the outermost for loop to a while loop instead:

find . -type d | while read D; do
    # ...
done

And btw, don't use `cmd` style command substitution, the modern way is $(cmd).

Deeply nested code can be hard to read. I recommend to skip the "." directory this way instead:

find . -type d | while read D; do
    [[ $D == . ]] && continue
    # ...
done

That is, instead of nesting deeper under an if, just revert the condition and do a continue instead.

And you don't need to quote ..

Instead of this:

folder_name=${D##*/}

Better to get the folder name like this:

folder_name=$(basename "$D")

Instead of this:

cd "$D"
# do something ...
cd ..

I would recommend this:

pushd "$D"
# do something ...
popd

The end result is the same, with the difference that you don't need to specify where to go back, it will go back to the right place. It's more robust that way.

You could get the PDF count much easier:

pdf_count=$(ls "$D"/*.pdf | wc -l)

This is a lot faster too, instead of iterating over each file and extracting their extensions, matching, and counting one by one.

You could refactor that part of the code to never go inside $D, which is more robust. (It's very easy to make mistakes when you change the working directory in the middle of scripts.)

Finally, most of the comments are pointless, the code you commented already explains itself, doesn't it?