Hi, the IndexHits instance returned isn't putting everything in that set
right away, but over time to avoid returning duplicates when combining
transaction state and store state. Looking at it right now I see that this
can be made much better by returning hits from transaction state first,
putting _only_ those ids into that set and then comparing - but not adding
- when iterating returning ids from store.
Let me see if I can get around fixing that soon...
On Friday, April 17, 2015 at 3:27:21 PM UTC+2, Rita wrote:
>
> I am very sorry to point out that those Lucene queries actually have a
> changeable behaviour, usually they are slower than the past, and in most
> case I get that error again (using 6GB heap of 8GB total RAM).
>
> I am updating the graph with delete and update of nodes and relationships.
> I decreased the number of operations per transaction but most of times I
> still got this error.
>
> The insertion with Batch Inserter instead seems to be ok!
>
> What do you suggest me please? Keep 1.9.9 version or upgrade to 2.2.1
> could be a solution? The new version covers packages related to these
> problems?
>
>
> Thanks in advance
>
> Rita
>
> Il giorno giovedì 16 aprile 2015 14:22:34 UTC+2, Rita ha scritto:
>>
>> Thank you for the reply Michael. I have just published the issue.
>> I was using as usual -Xmx4g. Now I've just tried with 6GB and open and
>> close a single transaction for every query like that on the different
>> indexes and I do not get this exception.
>> So now it has more need of memory for the same operation. I try on other
>> cases. Tell me if there are news on the issue please.
>> Thank you.
>>
>> Regards
>> Rita
>>
>> Il giorno giovedì 16 aprile 2015 13:04:01 UTC+2, Michael Hunger ha
>> scritto:
>>>
>>> This seems to be a like a bug.
>>>
>>> How much heap do you have?
>>>
>>> Could you raise an issue on github.com/neo4j/neo4j/issues ?
>>>
>>> Thanks so much
>>>
>>> Michael
>>>
>>> Am 16.04.2015 um 11:26 schrieb Rita <rita...@gmail.com>:
>>>
>>> Hi all,
>>> I am passing from Neo4j 1.9.9 to Neo4j 2.2.0, with embedded mode using
>>> java. I have inserted the transactions also for read operations, but now
>>> when I query my Lucene indexes as this
>>>
>>> rhits = index.query("cs", "*");
>>> out.println("#" + rhits.size());
>>> rhits.close();
>>>
>>>
>>> as you can see I do not have to iterate over the result, I need only the
>>> number of results but this new Neo4j version looks like loading all in
>>> memory and I get the following error in the first instruction.
>>>
>>> Exception in thread "main" java.lang.OutOfMemoryError: Java heap space
>>> at org.neo4j.collection.primitive.hopscotch.
>>> IntArrayBasedKeyTable.initia
>>> lizeTable(IntArrayBasedKeyTable.java:54)
>>> at org.neo4j.collection.primitive.hopscotch.
>>> IntArrayBasedKeyTable.<init>
>>> (IntArrayBasedKeyTable.java:48)
>>> at org.neo4j.collection.primitive.hopscotch.LongKeyTable.<init>(
>>> LongKeyT
>>> able.java:27)
>>> at org.neo4j.collection.primitive.Primitive.longSet(Primitive.
>>> java:66)
>>> at org.neo4j.kernel.impl.coreapi.LegacyIndexProxy$1.<init>(
>>> LegacyIndexPr
>>> oxy.java:296)
>>> at org.neo4j.kernel.impl.coreapi.LegacyIndexProxy.wrapIndexHits(
>>> LegacyIn
>>> dexProxy.java:294)
>>> at org.neo4j.kernel.impl.coreapi.LegacyIndexProxy.query(
>>> LegacyIndexProxy
>>> .java:352)
>>>
>>> I never get this with older versions of Neo4j! I always did this
>>> operation until version 1.9.9.
>>> Could you please help me to avoid this? Is it a bug of library
>>> implementation or I have to change the way of querying?
>>>
>>> Thanks in advance,
>>> Rita
>>>
>>> --
>>> You received this message because you are subscribed to the Google
>>> Groups "Neo4j" group.
>>> To unsubscribe from this group and stop receiving emails from it, send
>>> an email to neo4j+un...@googlegroups.com.
>>> For more options, visit https://groups.google.com/d/optout.
>>>
>>>
>>>
--
You received this message because you are subscribed to the Google Groups
"Neo4j" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to neo4j+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
