The use of the pipe separator is actually morte predictable than the original cpode suggests. It only ever occurs in leaf nodes and signifies that the datapoints are attributes within a datatype. I think you can dop a simple search/replace to prefix a / character.
The second fix was to handle multiple occurrence items i.e those with [1] indexer etc. The existing code was not handling these correctly. Have added a new test file which includes an example.
The use of the pipe separator is actually morte predictable than the original cpode suggests. It only ever occurs in leaf nodes and signifies that the datapoints are attributes within a datatype. I think you can dop a simple search/replace to prefix a / character.