removing strange characters from filenames

[ianholmes]

My application picks up external files from host systems.
I then run a script to reformat them, and transfer them to other servers.
I am finding that, on occasions, strange characters, which might be control characters, are included in the filename, which cause my script to misbehave.
eg A file might look like abc.def, but is actually $abc.$def.
How to I enhance my script to remove the "$" signs, which might be separate characters, or might be part of the "a" and the "d" (i.e. all one character )?
I know about "tr -d", but this doesn't seem to work for filenames, just file contents.
Many thanks

 

[jad]

echo 'thin\$g.\$ext' | tr -dc 'a-zA-Z\.'

should return 'thing.ext' if that helps

 

[aixinstall]

When i delete file's with control characters, i do the following,

make a listing of the directory in a file,
# ls > dirlist
Delete all the files you do not want to delete,
# vi dirlist
If only the files you want to delete are in dirlist then do
# rm `cat dirlist`

Good luck.....

 

[ShantanuB]

To furthur this dicussion use -- after a command.
ex, try this.

$ touch -- -Z
This will create a file with name "-Z"!!.
Now no matter what you do you can't rename it(mv), or edit it(vi,..) or even remove it(rm), since -Z will be read as an option instead of filename.
To get around this use -- again.
$ rm -- -Z # To delete the file
$ mv -- -Z Z.txt # To rename it ot Z.txt
etc etc..
In nutshell -- after a command tells shell command to take anything after it as it is.
Thus this can be used in your case.

 

[vgersh99]

rm ./-Z vlad
+---------------------------+
|#include<disclaimer.h> |
+---------------------------+