I have many files with duplicate names spread out over several tens of directories. I would like to mv them to the parent directory, but to avoid conflicting filenames I'd like to prefix each filename with the name of the directory it was in.

For example, if this is my directory structure:
/pictures/dotan/1.jpg
/pictures/dotan/2.jpg
/pictures/dotan/3.jpg
/pictures/ety/1.jpg
/pictures/ety/2.jpg
/pictures/ety/3.jpg

Then I'd like to end up with this:

/pictures/dotan-1.jpg
/pictures/dotan-2.jpg
/pictures/dotan-3.jpg
/pictures/ety-1.jpg
/pictures/ety-2.jpg
/pictures/ety-3.jpg

Assuming that pwd is /pictures, I figure that I'll need two expressions, one to rename and another to move. However, I don't know where to start. I've googled the cp (for renaming) and file (for going through all the directories recursively) however I don't know how to access the name of the directories. As for recusively moving the files afterwards, I will probably use file again.

A possibility for the given directory srtucture:

Try it first without the last pipe command to sh (in blue) to be sure it gives you the proper commands.


Regards

Thank you, that got me close enough to turn it into this:
However, now I've added yet another tool to my box! I've been avoiding awk for the time being, until I get a handle on more basiv stuff. Although I now have working code, I would still like this to be a positive learning experience. Is it possible to do this with regular bash tools: find, mv, grep, and so forth? I would really like to know how to do this with the commands that I'm currently learning, without getting into awk or perl.