bash extract first number from filename

Question

I'm trying to move files like this:

...
DLG023-00-01-sprite.TextGrid.json
DLG023-00-01.prosody.json
DLG023-01-01-sprite.ogg
DLG023-00-02-sprite.TextGrid.json
DLG023-00-02.prosody.json
DLG023-01-02-sprite.ogg
...
DLG028-00-01-sprite.TextGrid.json
DLG028-00-01.prosody.json
DLG028-01-01-sprite.ogg
...

into their respective folders, e.g.: move the file to dialogue023/ when it starts with DLG023; to dialogue024/ when it starts with DLG024; and so on.

All of the files and directories are at the same level.

In other words, I'm trying to produce commands like this one:

mv DLG023-00-01-sprite.ogg dialogue023/

How can I extract the first number from the filename?

So far I've got this:

for i in DLG*-*-*-sprite.ogg; do echo "${i//[^0-9]}";done

But this outputs all of the digits (e.g.: 230001) when I only want the first one (e.g.: 23).

Answer 1

I'm assuming that you are using OS X, which doesn't have the -t option in mv. This will be painfully slow if the script is processing thousands of files.

    for file in DLG*; do
    # Extract the three digits after DLG
    getnum=${file%%-*}
    getnum=${getnum#???}

    # make the directory dialogue### or exit if there is an error
    echo mkdir -p "dialogue${getnum}" || exit

    echo mv "$file" "dialogue${getnum}"
    done

Remove the echo before mkdir and mv if the output is satifactory.

Answer 2

You could generate the mv command using this:

for f in DLG*; do echo "$f" | sed 's/DLG\([0-9]\+\).*/mv "\0" "dialogue\1\/"/g'; done

Now, the command are printed like this:

mv "DLG023-00-01.prosody.json" "dialogue023/"
[...]
mv "DLG028-00-01.prosody.json" "dialogue028/"
[...]

If it is what you want, just pipe that output to bash and it will be executed:

for f in DLG*; do echo "$f" | sed 's/DLG\([0-9]\+\).*/mv "\0" "dialogue\1\/"/g'; done | bash

Answer 3

One line solution using find:

find . -name DLG* -exec bash -c 'DEST=$(echo "$0" | sed -r "s/DLG([0-9]+)-.*/dialogue\1/"); mkdir "$DEST"; mv "$0" "$DEST"' {} \;

• find command look for DLG* files from this directory.
• For every result execute the bash -c sentence.

In the bash sentence:

• DEST=$(echo "$0" | sed -r "s/DLG([0-9]+)-.*/dialogue\1/"); picks every result from find ($0) and using sed replace the all numbers before the first - (DLG([0-9]+)-.*) with dialogue following with those numbers (dialogue\1) and put in on a variable called DEST.
• mkdir "$DEST" creates the required folder.
• mv "$0" "$DEST"' moves the file to that directory.

Answer 4

A couple of ideas:

for i in DLG*-\*-\*-sprite.ogg; do echo "${i//[^0-9]}" | head -c 2; echo ;done

If you are sure they are all same length:

for i in DLG*-\*-\*-sprite.ogg; do echo "${i:3:3}" ;done