Removing numbers from a wordlist.

[oracx]

I would like to have the possibility to remove all entry’s that exists in file a from file b. If file a is numeric only and file b is alphanumeric, remove all numeric only entry’s from b.

[Archangel-Amael]

I moved your post from the previous thread, since it doesn't seem to fit with the threads theme.
As for removing numbers from a text file an easy way would be to use sed to do it.

root@bt~# sed 's/[0-9]*//g' input.txt
root@bt~# sed 's/[0-9]*//g' input.txt > output.txt

input.txt being the name of the file

[thorin]

You can use the translate command to remove digits.
The following example assumes that the file which contains the information needing the numbers removed is "input.txt" and the new file with the numbers removed will be "output.txt":

cat input.txt|tr -d [:digit:] > output.txt

tr = translate
-d = delete
[:digit:] = the thing you want to delete (this is tr specific syntax)

Try

tr --help

or

man tr

for more.

In your original post I'm a little confused why you mentioned files a and b if you're only interested in changing file b.

You can use

sort -u

for unique entries in files. i.e.:

cat *.txt|sort -u

Would find all the unique entries in text files for the current directory.

If file a is always a subset of file b you could sort the two then diff them.

You could also use grep or egrep and regex to identify numeric only strings and the -v switch for inverse matching.