Removing special chars from variable

[RageAx]

I have a variable that I need to remove special characters from to form a username. Since the variable is dynamic I cannot just simply replace the char in one position. I need to replace chars such as "čćšžđ" in upper and lower cases with their ISO counterparts such as "cszd". I have searched the command but I cannot find anything that would suit my needs.

[tikker]

You'll have to write a small program yourself. Store the letters with diacritics and their normal versions in two constants, say strings and use them to remap each character in the username to its normal counterpart is what I would try.

If you are on linux, then maybe the iconv command can help you; convert it from whatever encoding to plain ASCII.

[RageAx]

3 minutes ago, tikker said:

Store the letters with diacritics and their normal versions in two constants, say strings and use them to remap each character in the username to its normal counterpart is what I would try.

I did forget to mention I am a complete noob when it comes to Bash. So I didn't understand most of what you just wrote

[Neftex]

https://stackoverflow.com/questions/10207354/how-to-remove-all-of-the-diacritics-from-a-file

 

first answer uses some library, but someone said it didnt work correctly in some cases, the second answer seems like it would work correctly

[tikker]

Just now, RageAx said:

I did forget to mention I am a complete noob when it comes to Bash. So I didn't understand most of what you just wrote

Well you didn't mention any OS, environment or language so I gave a general programmatic strategy (didn't notice the tag). See my edited answer. Maybe iconv can help you.

[RageAx]

5 minutes ago, tikker said:

Well you didn't mention any OS, environment or language so I gave a general programmatic strategy (didn't notice the tag). See my edited answer. Maybe iconv can help you.

Can you use iconv on variables? I thought you can only use it on files.

And even if I use it on a file it justs returns '?' and not the respective character

[Neftex]

maybe your bash environment isnt set up with correct encoding so it cant display what youre converting to

[tikker]

34 minutes ago, RageAx said:

Can you use iconv on variables? I thought you can only use it on files.

And even if I use it on a file it justs returns '?' and not the respective character

You can make just pipe it in:

username_in="čćšžđ";
username_out=$(echo $username_in | iconv -f UTF-8 -t ASCII//TRANSLIT);
echo 'Input:' $username_in
echo 'Output: ' $username_out

which gives

Input: čćšžđ
Output:  ccszd

Replace UTF-8 with whatever encoding your usernames have. The //TRANSLIT part tries to approximate characters that aren't in the target encoding with similar looking characaters.

[Dat Guy]

27 minutes ago, tikker said:

The //TRANSLIT part tries to approximate characters that aren't in the target encoding with similar looking characaters.

Horrible idea in languages with similarly looking characters.

[tikker]

9 minutes ago, Dat Guy said:

Horrible idea in languages with similarly looking characters.

True, but you wouldn't want to represent those in ASCII then anway, but rather find a different way of handling them.