thanks Andreas: that is great... now we can publish a conformant nquads
reader.

I do have another question about the blank nodes. I want to do a
multistream load of the BTC but in that case I have to be sure that
blank nodes are completely local to named graphs. Do you know if this is
the case? To be more precise: would it be possible that an blank node
that was introduced in one named graph is referred to in another named
graph?

Jans

Andreas Harth wrote:
>
>
> Hi,
>
> I've updated the BTC 2009 dataset at [1]. The updates
> are minor fixes and concern encoding issues: Unicode
> escaping in URIs and bnode identifiers containing
> characters outside the alphanumeric range.
>
> .
>
>