Regex for reformatting a string to a safe xml-tag

Question

I'm currently trying to write a RegexReplace to ensure an input can be used as a valid XML-tag, meaning: no spaces, no special characters, only lowercase, etc...

Is there a common approach to that or do I have to do it all from scratch?

Example:

string Invalid = "asd(%4 asKUd n!%mn &§a_As1";  // Invalid as a tag

string Valid = FormatToSafeXmlTag(Invalid);  // How to write this function?

// Valid = "asd4_askud_nmna_as1"

Answer 1

• Only lowercase: ^[a-z]+$
• First char lowercase, (optional) remaining chars lowecase/numbers ^[a-z][a-z0-9]*$
• Only uppercase: ^[A-Z]+$
• First char alphabet, (optional) remaining chars alphanumeric ^[a-zA-Z][a-zA-Z0-9]*$

EDIT: To trim off everything but lowercase characters in javascript:

str = str.replace(/[^a-z]/g, "");

The catch is when users enters nothing but unacceptable characters - you will end up trying to create an xml tag with an empty string. I'd rather ask user to try again - how hard it can be to enter a lowercase string?

CAUTION: Another edge case is when user enters xml or any case-insensitive variants thereof (thanks to @Tim's answer). If you are on javascript, you cannot use the solution suggested by Tim, as it uses lookbehind, a feature unsupported by the javascript regex.

JavaScript code:

str = str.replace(/\s/g, "_"); //replaces spaces

str = str.replace(/[^a-zA-Z0-9_\-]/g, "");//trim symbols

var reg = new RegExp(/^xml/i); 

if(str.length == 0 || reg.test(str)) //is it empty or "xml" or "XmL" or ..
    alert("invalid regex");

Answer 2

According to the XML specification, an element's name is formed in the following way:

Name   ::=  NameStartChar (NameChar)*

Where

NameStartChar  ::=  ":" | [A-Z] | "_" | [a-z] | [#xC0-#xD6] | [#xD8-#xF6] 
  | [#xF8-#x2FF] | [#x370-#x37D] | [#x37F-#x1FFF] | [#x200C-#x200D] 
  | [#x2070-#x218F] | [#x2C00-#x2FEF] | [#x3001-#xD7FF] | [#xF900-#xFDCF] 
  | [#xFDF0-#xFFFD] | [#x10000-#xEFFFF]
NameChar       ::=  NameStartChar | "-" | "." | [0-9] | #xB7 
  | [#x0300-#x036F] | [#x203F-#x2040]

Which is trivial to convert to a regular expression.

If you're looking to remove any character outside of this definition, simply invert the characters the expression is looking for.

Answer 3

XML tags (I assume you're asking about tag names) have to follow these rules:

• start with letter, dot, colon or underscore
• only contain letters, numbers, dot, underscore or colon (for namespaces)
• must not start with xml

Therefore, a regex for valid tag names could be:

^(?!xml)[\w.:][\w\d.:]*$

depending on your regex flavor (e.g., .NET includes Unicode letters in \w, as is legal for a tag name). You could also use

^(?!xml)[p\{L}._:][\p{L}\p{N}._:]*$

if \w doesn't contain Unicode letters.

But of course you can use more restrictive rules, and possibly not all XML parsers can handle full Unicode tag names. So in the end,

^(?!xml)[A-Za-z._:][A-Za-z0-9._:]*$

might be your best bet...