Closed tpluscode closed 2 years ago
The data structure, as it is now, was probably aligned to the spec. The internal data structure will be adapted in an upcoming refactoring, as the parser supports only a single table schema and will do so in the future to avoid joins.
It would still be nice to be able to map multiple tables from a single CSVW.
I guess I was wrong and it would not require the use of groups. There can be multiple tables
in the single (implicit) group, thus allowing multiple output mappings from a single pass of a csv input. Wouldn't that be desirable?
I noticed that the parser only supports on schema per table but the property in
Metadata
class is an arrayhttps://github.com/rdf-ext/rdf-parser-csvw/blob/cae7d1d175f808079e3bf2aa8f450f83ceafe32a/lib/metadata/Metadata.js#L21-L25
On the other hand, the clearly defines
tableSchema
in section 5.4.2 as:Was the intention to extend the standard CSVW behaviour and support multiple mappings from on CSV? This would certainly be useful when producing multiple outputs from a single source file in one pass (what we currently do by iterating a single CSV multiple times when mapping multiple times)
Otherwise, would that feature require partially implementing table groups?