Parameters or arguments in bash - bash

Question

I saw a question on stackflow about parsing arguments. I tried to write this, but it's not working and now it's getting on my nerves.
The usual way of running a script on the terminal is ./scriptname, but I later introduced the argument -d. So, if I put ./scriptname it will not run. If I put ./scriptname -d it will.
Now I want to put another argument for the path (where the files are moving, in this case "/home/elg19/documents") such that when I do not include the path, it won't run. But, if I put ./scriptname -d path I want to replace $To in the existing script with the command argument after -d.
#!/bin/bash
From="/home/mark/doc"
To=$2
if [ $1 = -d ]; then
cd "$From"
for i in pdf txt doc; do
find . -type f -name "*.${i}" -exec mv "{}" "$To" \;
done
fi

Answer 1

Your desired usage isn't completely clear, but it seems to be:
scriptname -d path
So, you can do it the extensible way, or the brute force way. Since you're changing directories willy-nilly, you also need to ensure that the paths are absolute, not relative.
Brute force
#!/bin/bash
From="/home/mark/doc"
if [ $# = 2 ] && [ "$1" = '-d' ] && [ -d $2 ]
then
case "$2" in
(/*) cd "$From" &&
for extn in pdf txt doc
do find . -type f -name "*.$extn" -exec mv {} "$To" \;
done;;
(*) echo "$0: path name must be absolute ($2 is not)" 1>&2; exit 1;;
esac
else
echo "Usage: $0 -d /absolute/dirname" 1>&2; exit 1
fi
Extensible
#!/bin/bash
From="/home/mark/doc"
To=""
usage()
{
echo "Usage: $(basename $0 .sh) -d /absolute/dirname" 1>&2
exit 1
}
while getopts d: opt
do
case "$opt" in
(d) if [ ! -d "$OPTARG" ]
then echo "$0: $OPTARG is not a directory" 1>&2; exit 1
else
case "$OPTARG" in
(/*) To="$OPTARG";;
(*) echo "$0: path name must be absolute ($2 is not)" 1>&2; exit 1;;
esac
fi;;
(*) usage;;
esac
done
shift $(($OPTIND - 1))
if [ $# != 0 ] || [ -z "$To" ]
then usage
fi
cd "$From" &&
for extn in pdf txt doc
do find . -type f -name "*.$extn" -exec mv {} "$To" \;
done
For example, it will be very easy to add a -f from option to deal with changing the source of the files.
Note that you could also use:
for extn in pdf txt doc
do find "$From" -type f -name "*.$extn" -exec mv {} "$To" \;
done
This would allow you to permit relative names for the 'from' and 'to' directories because it does not change directory.

Answer 2

I assume you want to do some input validation to your command line arguments. I guess the following would be somewhat useful:
#!/bin/bash
usage() {
echo "USAGE :"
echo "./move -d <to-directory>"
}
if [ $# -ne 2 ] ; then
usage
exit
fi
case $1 in
-d ) shift
To=$1
;;
* ) usage
exit
esac
From="/tmp/From/"
cd "$From"
for i in pdf txt doc; do
find . -type f -name "*.${i}" -exec mv "{}" "$To" \;
done
Moreover to debug your script, you may use the following command:
bash -x ./move.sh -d /tmp/To/
You may add more error checking (and informative echo's) for the following cases:
Source/destination directory does not exits
N files have been copied from the to
No files available at
You can take the type of files as arguments f.e. -t doc xls pdf