Script to rename thousands of files and directories?

oh boy i don't even know where to begin with this one :(

OK here goes:
1)DO NOT use: for baz in $(command) !!
The reason is that the for loop while have to wait until the whole find command is done and store the whole freaking list in memory while it's waiting.
A more effective loop is :

command | while read baz; do stuff;done

2)DO NOT try to fight the shell ! Word splitting is there for a very good reason and bash has all the tools to work with it.

3)Study the possible input. What happens when a filename contains a newline? or a hyphen ? that's right, your script fails ! The right delimiter to use is $'\0' since filenames can't contain it. This is another reason not to use for loops, but while read -d $'\0' along with find's -print0 .

4)limit the input. why should your script operate on files that don't even contain spaces ? huh ?
find -name '* *' solves that

5)avoid using sed for such a trivial task. Use parameter expansion instead, like foo=${foo// /_}.man bash, section parameter expansion

6)help echo

good luck !

Great ideas for tweaks and modifications. Remember, though, that one of the key concepts for these sort of throwaway (e.g. "use once") scripts is that quick and dirty trumps elegant and optimized. If this were a script I'd use from a cron job every 15 minutes for the next six years I'd write it differently, but for a lot of people who work within the Unix or Linux environment, learning how to get it done and get on with their job is far, far more valuable than learning The One Best Way to do things. :-)

oh, quick and dirty ey ?

find . -type d -iname '* *' -exec sh -c 'mv "$1" "${1// /_}"' -- {} \;
find . -type f -iname '* *' -exec sh -c 'mv "$1" "${1// /_}"' -- {} \;

No need for a script then :P

Seriously, being correct is ALWAYS more important than being fast in the real world. If you learn the right way to write shell scripts, it WILL save time in the long run (whether it's a production machine or a desktop!). Trust me on this one.

Anyway, at least learn how word splitting works and why it's there, it'll pay back quickly, i promise ;)

How about handling filenames like:

Hey! Santa's Movie.mpg

Rather filenames with "'" in them.

eg:
Linux's_kernel.doc

thanks in advance..!!

meh, none of this appears to work right for:
/myfolder/First Spacer/somefolder/Second Spacer/

Quick and dirty should at least work fully the first time it's run not requiring multlple runs to make sure it gets everything.

Hello Dave !

In this example (and similar in others) You use
the loop :

for name in $(find . -type f -print | sed 's/ /____/g')
do
...
done

You get here after the substitution of the
find-command a very long line. In older shells
I made the experience, that 1000 words (in Your
example files) and in newer Shells 10000 is the
limit of then number of words for substitute.
To prevent such a performance-dependent loop, You
better write it in the following manner:

find . -type f -print | sed 's/ /____/g' | while read name
do
...
done

keep on doing and nice greetings from germany !

hi dave, i have been trying to write a script for mac OSX. i need to rename files eg :
116786_mcc_V1.jpeg to change to 116786.jpeg

can a script be written for this?

thanks
akbar

hi dave
i have been trying write a script for mac OSX
i need to rename a file eg
116786_mcc_v1.jpeg to 116786.jpeg

Can this script be written?

thanks
akbar

That's quite easy, Akbar. Look at the "for" loop, then figure out how to transform a current filename into a new filename. Something like this:

for name in *jpeg
do
newfilename = $(echo $name | transformation)
mv $name $newfilename
done

I'll let you figure out the transformation in the above. :-)