Re: Request for comments: Breaking down the datatype mapping problem (ISSUE-69)

On 21 Nov 2011, at 19:40, David McNeil wrote:
>> [[
>> Note: R2RML processor implementations are expected to augment the table with additional rows for mapping vendor-specific datatypes to appropriate RDF-compatible datatypes, like the XML Schema built-in types.
>> ]]
>> 
>> Is this an acceptable phrasing?
> 
> I think so. Truly I would have to spend more time puzzling through the spec to make sure I understand whether "supported SQL datatypes" includes vendor-specific datatypes.

I'm positive that it wouldn't require too much time nor puzzling ;-) See the note here:
http://www.w3.org/2001/sw/rdb2rdf/r2rml/#dfn-supported-sql-datatype

[[
The definition of supported SQL datatypes amounts to saying that any SQL datatype that can be cast to string by the SQL engine is supported. Casting to string is defined in SQL 2008 for all SQL datatypes except row types, array types, user-defined datatypes without a user-defined string CAST, and a few other exotic types. Vendor-specific types beyond SQL 2008 should be considered supported SQL datatypes, provided they can be cast to string by the SQL engine.
]]

> It seems that the "additional rows" you refer to would replace the final catch-all rule in the translation table for specific datatypes, right?

Well, in my mind they would be new rows in the augmented table, so the catch-all “Any other supported SQL datatype” would no longer apply to them in such an augmented table. This logic may not be too obvious – perhaps another sentence should be added to the Note to make that clearer.

Best,
Richard

Received on Monday, 21 November 2011 21:30:17 UTC