After improving the static XML harvester, and different kinds of XML records through, it revealed a flaw with the generic mapper. It should include element attributes through all node hops, not just the last one.
These lose the defining tag numbers becuase they are attributes before the terminating node, where currently only attributes are gathered from. If we collected them all the way down, the ES field names would be longer, but more meaningful. They would have been:
After improving the static XML harvester, and different kinds of XML records through, it revealed a flaw with the generic mapper. It should include element attributes through all node hops, not just the last one.
For example, the following MARC XML:
Results in the following fields:
These lose the defining tag numbers becuase they are attributes before the terminating node, where currently only attributes are gathered from. If we collected them all the way down, the ES field names would be longer, but more meaningful. They would have been:
However, it should probably not include the attributes form the root node, which often contains multiple declarations.