CODEWITHSUNDEEP
bash
user2846121
user2846121

Reputation: 13

Copying unique files to a new location

I am a total newbie to Linux and bash scripting and am currently stumped with this problem!

I have a directory containing many images from which I need to copy the unique images to a new location. I know there are numerous options for how to go about doing this but have very limited knowledge at the moment so appreciate I may be going about this the wrong way.

I used find and cat to create this list and have attempted to copy the files across with the intention of comparing them (using md5 and checking file names) when they are there.

However, the text file has 30 files on it but only 18 have been copied over. Can anyone advise?

My code to find files is -

find $1 -name "IMG_****.JPG" | cat > list.txt

and my code to copy from the list is

for image in $(cat list.txt);
do
cp $image $2
done

Upvotes: 1

Views: 1075

Answers (1)

slhck
slhck

Reputation: 38682

You're doing this much too complicated. Do not pipe find output to cat to pipe it into a list. This is an unnecessary use of cat. If you must, you can redirect the output of every program directly:

find "$1" -name "IMG_*.JPG" > list.txt

Also, do not use for to read lines from a file. Better use while with read:

while read -r filename; do
    cp "$filename" "$2"
done < list.txt

But it's even easier. You can just work with the files directly from find:

find "$1" -name "IMG_*.JPG" -exec cp {} "$2" \;

Here, {} will be replaced by each filename that find finds. Don't forget to quote your variables, so that spaces in file paths are no problem.

Another much simpler method with Bash options:

shopt -s nullglob globstar
cp -t "$2" -- "$1"/**/IMG_*.JPG

Here, globstar enables recursive matching of directories through **. The -t option to cp specifies the target of the copy operation.* The command will be expanded to cp -t target -- source1/IMG_foo.JPG source2/IMG_bar.JPG et cetera.

Now, as to your original issue, it could have been that some images have a space in their name. This would have broken your original script. If your image files contained a newline in their name, it also wouldn't have worked with while read … – but you would have gotten an error in that case of a file not being found.

Also note that cp overwrites files with the same name. Without asking for confirmation. So if in your subdirectories there are images with the same filename, you'd only get one result, with the latest overwriting the existing one.

* The -- isn't strictly necessary, but it's a good habit to include it to tell the command when the options arguments are over.

Upvotes: 6

Related Questions