Hi Sven,

how are you connected to the db (remote or plocal)?
Could you post the details of your db schema, in terms of classes,
properties and index definitions?

Thanks

Luigi


2015-04-21 17:03 GMT+02:00 Sven Hodapp <[email protected]>:

> Hi together,
>
> for my project I want to import a lot of data into the database. But the
> data should be de-duplicated.
>
> First of all, without de-duplication the insertion takes about 500 ms (a
> minimal test set):
>
>     docelem = graph.addVertex("class:DocElem", "uri", uri, "type", type,
> "model", model);
>
>     docelem.setProperties(attrs);
>
>     graph.commit();
>
>
> Now I *don't* want that the is uploaded twice, so I'll check it like
> this, if it's already in the database:
>
>
>     Iterable<Vertex> iter = graph.query()
>
>         .has("uri", Compare.EQUAL, uri)
>
>         .limit(1)
>
>         .vertices();
>
>
> Then, with iter.iterator().hasNext() I'm checking, if the vertex is
> already in the database. But this is dead slow (even indexed, or I've made
> a mistake)! Now it takes about 15 s for inserting.
>
>
> You can suggest a better solution? The best case would be, if I don't have
> to call the database; and the database recognizes that the requested uri is
> already inserted and may only update the entry, or something like that!
>
>
> Note 1: With println instead of db-insert the code needs about 50 ms to
> fetch/create the data. Is is possible to go faster?
>
> Note 2: I'm using OrientDB 2.1-rc1 with remote connection (on the same
> host).
>
>
> Regards,
>
> Sven
>
> --
>
> ---
> You received this message because you are subscribed to the Google Groups
> "OrientDB" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> For more options, visit https://groups.google.com/d/optout.
>

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to