Encode local name like XmlConvert.EncodeLocalName in pure XQuery

Question

Hello I'd like handle Xml-Files which have encoded node names like for example:

<CST_x002F_SOMETHING>
....
</CST_x002F_SOMETHING>

This node name should be decoded to CST/SOMETHING.

These node names were encoded for example via EncodeName. Is there any built-in XQuery-function to decode these names? Or do you have an encoding / decoding function?

XML Files produced by Oracle-DB use the same escaping mechanism.

Answer 1

Use fn:analyze-string() to split the string and match the _XXXX_ parts. When you encounter one of these parts, use bin:hex() to convert hex to binary, then bin:unpack-unsigned-integer() to convert the binary to an integer, then fn:codepoints-to-string() to convert the integer codepoint to a string.

The binary functions are documented at https://www.saxonica.com/documentation/index.html#!functions/expath-binary

Requires Saxon-PE or higher.

You could also use the new saxon:replace-with() function:

declare namespace bin = 'http://expath.org/ns/binary'; 
saxon:replace-with('CST_x002F_SOMETHING', '_x[0-9A-F]{4}_', 
   function($s) {$s => substring(3, 4) 
                    => bin:hex() 
                    => bin:unpack-unsigned-integer(0,2) 
                    => codepoints-to-string()} 

outputs CST/SOMETHING

Answer 2

You can try the following.

XQuery

for $x in doc('input.xml')//CST_x002F_SOMETHING
return rename node $x as "whateveryouneed"