Re: Proposal for ISSUE-32, curly braces in joinCondition

>
> Just curious, and apologies if I'm missing the context from weekly
> discussions that may have covered this, but isn't there also the issue of
> the various *types* of joins to consider?
>

Bob - That is a good point. Previously the working group agreed that only
inner joins made sense for this. The reason given was that you wouldn't
produce a triple for a null value. The joinCondition is used to join to
another TripleMap's logical table for the purpose of referencing it's
subject. Since the child triple cannot have either a NULL subject or a NULL
object there is no need to support outer joins.

This also raises the question of how NULL columns are handled in general by
R2RML. As far as I can tell the current draft of the recommendation does not
address this (although it does address the case of NULL values with the
graphColumn property). It seems to me that the spec should explicitly state
how NULL columns are to be handled in the case of the rr:column and
rr:template properties. Possible answers to this question are:

1) throw an error if a column is null
2) do not produce a triple if a referenced column is null
3) ??

-David