All exports from vahydro need consistent, generic dh format that preserves native data table names

[rburghol]

Changes in name for columns like "Source.Type" instead of "ftype" need to go so that we can leverage the consistency of the underlying structure from vahydro. This will not only make our code more reusable, it will also eliminate painful things like having to quote mixed case column names and column names with dots. If we want labels, we should export relevant dh_variabledefinition entries and match varname or varabbrev columns. Timeseries entries should therefore be exported with both varkey and varid joined in, and use featureid and entity_type. Likewise, propvalue and tsvalue should be the numerical column name in all exports.