Indeed, that works.

Maybe it's just me being a little dumb. Thanks!

On Thu, May 21, 2015 at 12:15 PM, Marko Rodriguez <[email protected]>
wrote:

> Hello,
>
> Did you iterate your traversal?
>
>         graph.tx().open(); g.addV().iterate(); graph.tx().commit();
>
> Marko.
>
> http://markorodriguez.com
>
> On May 21, 2015, at 10:00 AM, Dylan Millikin <[email protected]>
> wrote:
>
> > Hey,
> >
> > This one is somewhat obvious but a little bit of a pain and worth
> > discussing.
> > I'm using the titan09 branch and gremlin-console M9-RC3 to test these
> > (default configuration)
> >
> > Basically these two queries don't yield the same results:
> >
> > graph.tx().open(); graph.addVertex(); graph.tx().commit();
> >
> > Which will behave as expected and add a vertex
> >
> > &&
> >
> > graph.tx().open(); g.addV(); graph.tx().commit();
> >
> > Doesn't add any vertex. I understand why, I'm just not sure this is
> > comfortable behavior (or even wanted behavior).
> > Furthermore the following doesn't work either:
> >
> > graph.tx().open(); m = graph.traversal(); m.addV(); graph.tx().commit();
> >
> > So, as handy as addV() is, it's behavior is inconsistent and dependent on
> > context. That query you wrote and that worked fine will suddenly stop
> > working if it's called from within a transaction. (unit tests will pass
> but
> > integration tests will fail)
> >
> > Has this been discussed already? It looks like a potential headache.
>
>

Reply via email to