problem in script that change files

nadavvin · 12-29-2007, 11:06 AM

Hello

My friend ask me about script of changing large files with sub directories with index of them like:

01 file1
02 file2

I try to do one but the "find" command when their files with spaces return to the for loop each part.

for example file like "hello world.txt" will split to "hello" and "world.txt"

How do I use in for in bash loop correctly?

If there is script like that you can tell me about it

Code:

#!/bin/bash
THIS=`basename "$0"`
NUMBER=0
for file in $( find -type f ); do
	echo start loop
	echo $file
	FILENAME=`basename $file`
	if [ "$FILENAME" = "" ]; then
		echo problem
		exit
	fi
	if [ "$FILENAME" = "$THIS" ]; then
		continue
	fi
	DEST="`dirname $file`/$NUMBER $FILENAME"
	echo "$DEST"
	mv "$file" "$DEST"
	let NUMBER+=1
done

David1357 · 12-29-2007, 12:12 PM

Quote:

Originally Posted by nadavvin

Code:

	# Maybe change this
	FILENAME=`basename $file`

	# ... to this
	FILENAME="`basename $file`"

I think your version only assigns the everything up to the first space to FILENAME. I just tested this theory on the command line in bash.

radoulov · 12-29-2007, 12:14 PM

You could try something like this:

Code:

find -type f|while IFS= read -r;do
	mv -v "$REPLY" "${REPLY%/*}/$(printf "%0.2d" $i) ${REPLY##*/}"
	i=$((++i))
done

If you put it in a script and you want to exclude
the script name from the result, you should modify
the command like this:

[assuming GNU find]

Code:

find ! -wholename "$0" -type f ...

jschiwal · 12-29-2007, 12:35 PM

Code:

for file in $( find -type f ); do
    echo start loop
    echo "$file";
    FILENAME=`basename "$file"`
..

Use double quotes to preserve spaces. Including in `basename "$file"` and when referencing $FILENAME.
However, why are you messing with all of these basename and dirname functions.

Code:

mkdir test
jschiwal@hpamd64:~> touch test/"01 file"
jschiwal@hpamd64:~> touch test/"02 file"
jschiwal@hpamd64:~> touch test/"03 file"
jschiwal@hpamd64:~> mkdir destdir
jschiwal@hpamd64:~> for file in test/*; do
> mv "$file" destdir/ -v
> done
`test/01 file' -> `destdir/01 file'
`test/02 file' -> `destdir/02 file'
`test/03 file' -> `destdir/03 file'

Code:

find test/ -type f -print0 | xargs -0 mv -v --target-directory=destdir/
 `test/02 file' -> `destdir/02 file'
 `test/03 file' -> `destdir/03 file'
 `test/01 file' -> `destdir/01 file'

ghostdog74 · 12-30-2007, 09:46 AM

Quote:

Originally Posted by nadavvin

Hello

My friend ask me about script of changing large files with sub directories with index of them like:

01 file1
02 file2

I try to do one but the "find" command when their files with spaces return to the for loop each part.

for example file like "hello world.txt" will split to "hello" and "world.txt"

How do I use in for in bash loop correctly?

If there is script like that you can tell me about it

Code:

#!/bin/bash
THIS=`basename "$0"`
NUMBER=0
for file in $( find -type f ); do
	echo start loop
	echo $file
	FILENAME=`basename $file`
	if [ "$FILENAME" = "" ]; then
		echo problem
		exit
	fi
	if [ "$FILENAME" = "$THIS" ]; then
		continue
	fi
	DEST="`dirname $file`/$NUMBER $FILENAME"
	echo "$DEST"
	mv "$file" "$DEST"
	let NUMBER+=1
done

use while loop instead

Code:

...
find -type f | while read file
do
 ...#your code##
done
...

matthewg42 · 12-30-2007, 12:00 PM

An alternative to calling basename is to use the ${variable##pattern} expansion. This removes the longest string matching pattern from the value of variable, thus these two commands will output the same thing:

Code:

echo base name is `basename "$0"`
echo base name is ${0##*/}

Calling basename is less desirable because it invokes an external program, which means it is slower and less reliable. It's not a big deal, but it's a nice touch to use the shell expansions like this.