Closed srowen closed 3 years ago
@HyukjinKwon not sure if you'd like to review this - not necessary. My main concern is what I commented on at https://github.com/databricks/spark-xml/issues/480#issuecomment-706338172
I'm going to add this. It won't break existing workloads, and if the functionality works oddly in some cases mentioned above, well, it's still good to have it mostly working.
See https://github.com/databricks/spark-xml/issues/480
This attempts to support a special 'wildcard' schema element, called by default "xs_any", that will match any unmatched XML elements in a node, and treat them as strings containing the raw XML content. If it's a string, it will match one element. If it's an array of strings, it will match all of them.
This provides a natural interpretation of "xs:any" in XSDs then.