Major Description
This change ignores CDATA within row tags as well as outside of it. We should only ignore CDATA found outside of row tags as they are considered data within the row.
https://github.com/apache/spark/pull/45487
NOTE: With the current parser implementation, after not ignoring CDATA elements within row tags there remains the edge case of a matching closing row tag within CDATA which will be parsed as a valid end tag.
Example:
<row> <![CDATA[ </row> ]]>
after no longer ignoring CDATA within rows, the closing tag in the example above will be matched by the parser which is incorrect.