Hi,
This is fixed in the development builds.
We also know that loading into a named graph, after the fix, creates a
larger dataset than TDB1 would.
It is the same size as TDB1 if TDB1 used the same copy-style loader - for
some reason, TDB1 bulk loader does things in an order so that the indexes
are better packed.
There isn't a full TDB2 bulkloader yet. The command exists so the syntax of
use will remain but the algorithm is a naive one ATM.
> g:mygraph
Beware that graph name can't be a prefixed name, it is treated as URI.
Andy
On 25 November 2017 at 14:30, Laura Morales <[email protected]> wrote:
> I'm experimenting with the new TDB2 but I get some errors. I've copied the
> config file described in https://jena.apache.org/documentation/tdb2/tdb2_
> fuseki.html into "run/config.ttl". This seems to work fine, because I can
> start the endpoint and see the empty dataset. The problems are when
> creating the TDB2 store.
> I've downloaded Fuseki, then downloaded Jena inside the Fuseki root
> directory.
>
> $ ./jena-3.5.0/bin/tdb2.tdbloader --tdb run/config.ttl --desc
> run/config.ttl --verbose file.nt
>
> 15:22:08 WARN ModTDBDataset :: Unexpected: Not a TDB2 dataset for
> type DatasetTDB2
>
> ^--- I get this warning but otherwise the graph loads fine (as far as I
> can tell).
>
> $ ./jena-3.5.0/bin/tdb2.tdbloader --tdb run/config.ttl --desc
> run/config.ttl --graph g:mygraph --verbose file.nt
>
> 15:22:08 WARN ModTDBDataset :: Unexpected: Not a TDB2 dataset for
> type DatasetTDB2
> java.lang.ClassCastException: org.apache.jena.tdb2.store.GraphViewSwitchable
> cannot be cast to org.apache.jena.tdb2.store.GraphTDB
> at tdb2.cmdline.CmdTDBGraph.getGraph(CmdTDBGraph.java:70)
> at tdb2.tdbloader.loadNamedGraph(tdbloader.java:123)
> at tdb2.tdbloader.exec(tdbloader.java:113)
> at jena.cmd.CmdMain.mainMethod(CmdMain.java:93)
> at jena.cmd.CmdMain.mainRun(CmdMain.java:58)
> at jena.cmd.CmdMain.mainRun(CmdMain.java:45)
> at tdb2.tdbloader.main(tdbloader.java:54)
>
> ^--- Here I can't load the graph at all.
>
