I found that in a case when only a subset of columns are actually mapped, the default columns would not be output but instead a "random" predicate from the those mapped would be taken
This is because TableSchema#root was never actually set and thus this.dataset.match was called with a null parameter and returned an unexpected csvw:propertyUrl from whatever would be found.
The result would be similar to the below, where <http://schema.org/name> is used for all columns
I found that in a case when only a subset of columns are actually mapped, the default columns would not be output but instead a "random" predicate from the those mapped would be taken
This is because
TableSchema#root
was never actually set and thusthis.dataset.match
was called with anull
parameter and returned an unexpectedcsvw:propertyUrl
from whatever would be found.The result would be similar to the below, where
<http://schema.org/name>
is used for all columns