The error you're encountering is due to an invalid XML character reference in your string. Let's break down what's happening:

The Problem

• The string b��c contains two XML character references: � and �.
• These references correspond to Unicode code points U+D83D and U+DCA7, which are surrogate pairs used in UTF-16 encoding to represent characters outside the Basic Multilingual Plane (BMP), such as emojis.
• In XML, character references must refer to valid Unicode code points. U+D83D is a high surrogate and U+DCA7 is a low surrogate. Surrogate pairs are not valid on their own in XML; they must be combined to form a single Unicode character (e.g., the emoji 💥).

Why the Error?

• XML parsers expect character references to refer to valid Unicode scalar values (i.e., code points that are not surrogate halves).
• The parser sees � (U+D83D) and � (U+DCA7) as two separate, invalid references, hence the error: xmlParseCharRef: invalid xmlChar value 55357

How to Fix It

You have two options:

1. Use the actual Unicode character (emoji) directly in your XML:

xml_string = '''<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<hierarchy a="b💥c"/>'''
etree.fromstring(xml_string.encode('utf-8'))

This works because the emoji is a single Unicode character.

2. Use the combined code point for the emoji:

The emoji 💥 is actually U+1F4A5, which is the combination of the surrogate pair. You can use this directly in your XML:

xml_string = '''<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<hierarchy a="b&#128165;c"/>'''
etree.fromstring(xml_string.encode('utf-8'))

Here, 💥 is the decimal representation of U+1F4A5.

Summary Table

Why This Matters

• XML parsers do not handle surrogate pairs as separate entities; they must be combined into a single valid Unicode character.
• Always use the actual Unicode code point for characters outside the BMP, or use the character itself if your encoding supports it (UTF-8 does).