Still dealing with this problem as I mentioned earlier the slowdown just
gets longer and longer the more records I stuff in. But after the
application runs and exists I then open the DB and queries are very fast. I
also notice that while I am stuffing in records the size of the DB grows
pretty dramatically (4gb) but then once import is complete this size
shrinks down to around 200mb.
I am wondering if there is something I need to periodically do like tell
the DB to compact itself and update indexes or something? Its either that
or I guess I close the DB then re-open it to start adding more for every X
thousand vertices I insert. This seems like something I shouldnt have to
do...
On Tuesday, July 22, 2014 10:10:15 PM UTC-4, Seamus Minogue wrote:
>
> Yes that seems to be it. The time seems to increase as more records are
> inserted. When its <500 records I'm seeing like 10 per second. But when
> There are like 3500 records having been inserted I'm seeing as much as 3
> seconds per query.
>
> Do I need to do something to force the index to "re-index" the records
> being added?
>
> -Seamus
>
>
> On Tuesday, July 22, 2014 6:19:47 PM UTC-4, Lvc@ wrote:
>>
>> Hi Seamus,
>> So where most the time is spent? In the following line of code?
>>
>> for (Vertex v : graph.query().has("id", dbuser.id).limit(1).vertices())
>>
>> Lvc@
>>
>>
>>
>> On 22 July 2014 22:10, Seamus Minogue <[email protected]> wrote:
>>
>>> The scenario is that I am trying to pull a bunch of data out of an
>>> existing database and storing it in an orientDb which is running embedded
>>> in my application. I am running into some performance problems,
>>> specifically when I add queries to verify that the data I am about to
>>> insert into orient is unique the performance is killing me.
>>>
>>> I am only dealing with 7000 rows of data in my database as a test case,
>>> The below snippet of code is showing what I am doing. Assume that of the
>>> 7000 rows there are 1500 unique user ID's. When I just do inserts and dont
>>> worry about duplicates This runs in 7 seconds. When I add the query its
>>> taking ~20mins.
>>>
>>> I was going through the performance tuning suggestions trying to find an
>>> answer... I figure I must be doing something wrong.
>>>
>>>
>>> OServer server = OServerMain.create();
>>>
>>> server.startup(...);
>>>
>>> server.activate();
>>> graph = new OrientGraphNoTx("plocal:/tmp/test", "admin", "admin");
>>>
>>> graph.getRawGraph().declareIntent(new OIntentMassiveInsert());
>>> graph.getRawGraph().setValidationEnabled(false);
>>>
>>> graph.getRawGraph().setRetainRecords(false);
>>>
>>> graph.createKeyIndex("id", Vertex.class, new Parameter<String, String>(
>>> "type", "UNIQUE"), new Parameter<String, String>("class", "User"));
>>>
>>> //Read user from existing store and create object (mundane details)
>>> User dbuser = fetchUserFromOutside();
>>>
>>> Vertex user = null;
>>>
>>> for (Vertex v : graph.query().has("id", dbuser.id).limit(1).vertices())
>>>
>>> {
>>> user = v;
>>> }
>>>
>>> if(user == null){
>>>
>>> user =graph.addVertex("class:User","id", dbuser.id);
>>>
>>> }
>>>
>>>
>>>
>>> --
>>>
>>> ---
>>> 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.
