Create folders from space-delimited filenames and copy files into them

Question

I'm currently trying to organize several thousand files which are named according to what's in them, and the various "tags," if you will, are separated by spaces. So, for example:

foo_bar bar_foo.txt

I'm relatively new to Linux/Unix, and was wondering if there was a way to iterate through every file, create folders based on the tags, and copy the files to those folders?

So we'd end up with:

./foo_bar bar_foo.txt
./foo_bar/foo_bar bar_foo.txt
./bar_foo/foo_bar bar_foo.txt

So far, I've been manually doing everything like this:

mkdir foo_bar
cp *foo_bar* foo_bar/
mkdir bar_foo
cp *bar_foo* bar_foo/
...

Obviously this is pretty time-inefficient, so I'm just looking for a way to automatically do it.

Edit: Some more examples:

Input:

./a b c d.txt
./b a d.txt
./c d e.txt
./d a.txt

Output:

All original files still in parent directory, plus:

./a/a b c d.txt
./a/b a d.txt
./a/d a.txt

./b/a b c d.txt
./b/b a d.txt

./c/a b c d.txt
./c/c d e.txt

./d/a b c d.txt
./d/b a d.txt
./d/c d e.txt
./d/d a.txt

./e/c d e.txt

Answer 1

try this.. remove the echo.. if you are happy with the commands...

ls * | awk '{print $1}' | sort -u | while read a; do echo mkdir -p $a; echo mv ${a}*txt ${a}; done

modified answer

$ ls
    a b c d.txt  b a d.txt  c d e.txt  d a.txt
$ ls * | sed "s/.txt//;s/ /\n/g" | sort -u | while read file; do echo mkdir -p $file; echo mv *${file}*.txt ${file}; done
    mkdir -p a
    mv a b c d.txt b a d.txt d a.txt a
    mkdir -p b
    mv a b c d.txt b a d.txt b
    mkdir -p c
    mv a b c d.txt c d e.txt c
    mkdir -p d
    mv a b c d.txt b a d.txt c d e.txt d a.txt d
    mkdir -p e
    mv c d e.txt e

Note: as you want the file be present in your original directory.. use cp instead of mv