cdata-hack

When writing EJB-QL queries inside ejb-jar.xml, comparison operators like < and > conflict with XML’s own tag delimiters. The XML parser will try to interpret basePrice > 200 as XML tags, causing parse errors and deployment failure.

The CDATA (Character Data) section tells the XML parser to treat everything inside as raw text, not XML markup. It is a way to “escape” special characters without escaping each one individually.

1. XML normally interprets < as the start of a tag and > as the end of a tag
2. Wrapping content in <![CDATA[ ... ]]> disables XML parsing for that section
3. In EJB-QL: SELECT OBJECT(p) FROM PRODUCTS p WHERE p.basePrice <![CDATA[>]]> 200
4. The parser reads the comparison operator as literal text, not markup
5. This is a generic XML technique, not specific to EJB — used anywhere XML contains code or special characters

• Syntax: <![CDATA[ content ]]> — brackets and parentheses are mandatory
• Content inside is treated as raw character data
• Only solution for using >, <, >=, <= inside XML element text
• Not needed in Java code or non-XML contexts

• Built from: EJB Deployment Descriptor — CDATA is used inside XML deployment descriptors
• Builds into: EJB-QL — EJB-QL queries in XML need CDATA for operators
• Related: CMP — CMP uses EJB-QL in deployment descriptors
• Related: XML Basics — general XML escaping concept

• ]]> cannot appear inside a CDATA section (it would end the CDATA prematurely)
• CDATA is not the same as XML entity escaping (< and >) — both work but CDATA is cleaner for large code blocks
• Forgetting CDATA causes cryptic XML parse errors during EJB deployment