Moving files with a space

fritz1968 · 02-03-2015, 05:27 PM

I have a directory structure that at least 10 directories (same level). Each directory contains a mix of media files: *.jpg, *.JPG, *.AVI, *.MOV, etc...

I want to move certain files to a new path, preserving the directory structure. I have the following, but the problem is that I have several files with a space in them. THose files are not moved. Here is my bash code so far:

#/bin/bash
#######################################

source1=$1
destDir=$2
fileExt=$3

echo $source1
echo $destDir
echo $fileExt

cd $source1 # This is where the file were downloaded to

for file in $(find -name "$fileExt" ) #-type f -not -path "./.svn/*")
do
fPath="$destDir/$file" # This is the full path of the file to copy to
dPath=$(dirname ${fPath}) # This finds the directory of the path

if [[ ! -d "$dPath" ]] # double checking that the directory exists
then # It does not, so create it
mkdir -p $dPath
fi
echo "=================================="
echo Path : "$fPath"
echo File : "$file"
mv "$file" $dPath
done

The error for a file with a space in it (IMG_5210 - Version 2.jpg) is :
Path : ../testDen/./HolidayCardPics/IMG_5210
File : ./HolidayCardPics/IMG_5210
mv: cannot stat ‘./HolidayCardPics/IMG_5210’: No such file or directory
==================================
Path : ../testDen/-
File : -
mv: cannot stat ‘-’: No such file or directory
==================================
Path : ../testDen/Version
File : Version
mv: cannot stat ‘Version’: No such file or directory
==================================
Path : ../testDen/2.jpg
File : 2.jpg
mv: cannot stat ‘2.jpg’: No such file or directory

How do I correct my code so that it does not look at the spaces as different files?

Thanks in advance for your help.

mina86 · 02-03-2015, 05:53 PM

Use [code] around your code snippets for better readability.

Quote:

Originally Posted by fritz1968

for file in $(find -name "$fileExt" )

Unfortunately, no, this won’t work. You can try doing this:

Code:

find -name "$fileExt" | while read file; do

but even that will fail if any of the file contains a new line character. If I ever need to use find in such a way, I end up implementing it’s functionality in shell myself, i.e. something along the lines of:

Code:

handle_dir() {
    for entry in "$1/"*; do
        if [ -d "$entry" ]; then
            handle_dir "$entry"
        elif [ -f "$entry" ]; then
            handle_file "$entry"
        else
            echo "$entry: not a regular file or directory, skipping" >&2
        fi
    done
}

hanadle_file() {
    : … do what you need …
}

PS. Quote variables consistently and use ‘--’ to separate arguments from file names. For example:

Code:

mv -- "$file" "$dPath"
# NOT: mv "$file" "$dPath"

dugan · 02-03-2015, 06:15 PM

Use recursive globbing.

Code:

#/bin/bash
#######################################

shopt -s globstar

source1=$1
destDir=$2
fileExt=$3

echo $source1
echo $destDir
echo $fileExt

cd "$source1" # This is where the file were downloaded to

for file in **/$fileExt
do
	fPath="$destDir/$file" # This is the full path of the file to copy to
	dPath="$( dirname "${fPath}" )" # This finds the directory of the path

	if [[ ! -d "$dPath" ]] # double checking that the directory exists
	then # It does not, so create it
		mkdir -p "$dPath"
	fi
	echo "=================================="
	echo Path : "$fPath"
	echo File : "$file"
	mv -- "$file" "$dPath"
done

Also note the extra quoting (which was necessary).

grail · 02-03-2015, 11:46 PM

Another quick alternative for getting a directory path:

Code:

dPath="${fPath%/*}"