sed nightmare

[drakonite]

I'm trying to write a "simple" line to rename a bunch of files from things like "foo [T+meh] (J).bar" or "ick [o37] (JU).ack" into nice names like "foo.bar" and "ick.ack"

I've gotten this so far :

Code: Select all

for i in *;do mv \'$i\' \'`echo $i|sed -e "s/'/\'/g" -e 's/ \[[A-Za-z0-9\+\!]*\]//g' -e 's/ ([J U E])//g'`\';done

If I replace mv with echo what it prints out looks like it should be right.. but when running it with mv it complains "mv: when moving multiple files, last argument must be a directory" for all the files...

Can anyone help me sort out this sed nightmare?

[TGL]

The nightmare is probably not in the sed command, but in your filename quoting. I would have used:

Code: Select all

... ; do mv "$i" "`...`" ; ...

That said, I've not checked your sed command, but according to your examples, all you have to do is to remove in the names everything from the first space (included) to the dot (not included). I would do it that way:

Code: Select all

for i in * ; do mv "$i" "`echo $i | sed -e 's:\ .*\.:\.:'`" ; done

Or are there some other files with space in their name that you don't want to change? In that case, you can try something like this:

Code: Select all

for i in * ; do (echo $i | grep "\[[A-Za-z0-9\+\!]*\]\ ([JUE]*)") && mv "$i" "`echo $i | sed -e 's:\ .*\.:\.:'`" ; done

(which will also give you feedback about what files are being renamed)

[fyerk]

Is it a requirement that you use sed? It seems like this would be much easier as a combination of find and a Perl one-liner.

This should do what you need...

Code: Select all

find . -maxdepth 1 -type f | perl -ple '$a=$_;s/([a-z]+)\s\[[^\]]+\]\([^\)]+\)/\1/;rename $a, $_'

Explanation can follow in another post if you're interested...

[TGL]

It seems like this would be much easier as a combination of find and a Perl one-liner.

I tend to think that the right tool is the one you know, but it's somehow offtopic

[drakonite]

It seems like this would be much easier as a combination of find and a Perl one-liner.

I tend to think that the right tool is the one you know, but it's somehow offtopic

Yes, I have to agree I don't know perl (or don't remember anything useful atleast ) so my very trivial sed skills were what I chose.

Thanks for the help TGL. Changing how I was quoting did the trick. I've always been in the habit of using ' instead of " when typing filesnames.. hrmm, guess I should switch

Now I just need to make it check if it should overwrite files or not and I'll be set Shouldn't take me too long...

[pjp]

Moved from Other Things Gentoo.

[funkmankey]

I chose the lazier route and installed krename. yes, it requires kdelibs and yes, kde is pure evil, but I'd already had to install evil before for some reason or other (I forget why now, though), so it wasn't such a big deal after all.

yes, I have a pretty decent grasp of regexes. but it always seems to turn out wrong, with the salt and the pain and the burning and the lost filesystems...

[TGL]

I chose the lazier route and installed krename.

For command line, you also have "ren" (emerge sys-apps/ren) which is very good for this kind of task. Here, it would have something like this:

Code: Select all

# ren "* [*] (*).*" "#1.#4"

Depending on some options, it will or not be interactive. But it so much funnier with bash and sed (or perl ).

[drakonite]

so NOW you break in with the ren and krename programs... lol

[TGL]

Sincerely, I've only reminded "ren" while reading funkmankey's post about "krename". But it doesn't mean I would have talked you about it before you've fixed your sed version