Beautiful.
Thank you very much :)
Jon
On Sep 25, 2011 5:48 AM, Peter Hunsberger [via Neo4j Community
Discussions] ml-node+s438527n3365885...@n3.nabble.com wrote:
I'm going to take a slightly different tack here than the responses you've
got so far...
First as others have pointed out, this
hi McKinley,
This is my test code:
http://neo4j-community-discussions.438527.n3.nabble.com/file/n3366449/Entity.java
Entity.java
http://neo4j-community-discussions.438527.n3.nabble.com/file/n3366449/EntityTest2.java
EntityTest2.java
Do the testCreateNode to create a node with a property
If you've got 100 concurrent threads writing to one node only one can write
to it at a time. Write locks are held from the time you do a write operation
on that node until that transaction which made that write operation
finishes. So in this scenario your setProperty calls will be serial instead
I've executed the query in the cypher console. (Btw I had to turn
FireBug ON before it accepted the double enters)
START n=(1) match a--n where a.Gemeente='Antwerpen' return a
a.Gemeente does not exist on Node[891]
If I add the Gemeente-property on Node[891] and run the query again, he
Hi Filip,
if you need only those nodes who have the Geemente property, you can specify
it on the where clause. Example:
where a.Geemente and a.Geemente = 'Antwerpen'
with this, the verification for Antwerpen (lovely city, BTW) is only
executed if Geemente property exists on a node.
See if this
yeah,this is a real problem in my scenario. The 100 concurrent threads is
getProperty not setProperty.In my scenario I just sort a node list by node's
property. And I found it is not fast in concurrent requests.Then I found
these.
--
View this message in context:
Hi,
I am just trying to clarify my understanding of read and write locks. There
has been discussions around getting a write lock on a given node by removing
a non-existent property from the node, which in turn triggers
a nodeManager.acquireLock which calls lockManager.getWriteLock.
The only
Is anyone working on a bridge between WEKA and neo4j?
___
Neo4j mailing list
User@lists.neo4j.org
https://lists.neo4j.org/mailman/listinfo/user
You have some some things in your code that make it difficult to know what
you want to test. You are running 100 concurrent threads. However, each of
those threads is getting the property value of only one node. Each thread is
reading that single node's property one million times sequentially.
Bryce,
The situation you are facing seems to only be a problem with choosing the
type of the interface for your variable instead of the type that implements
the getConfig method. For example:
EmbeddedGraphDatabase graphDb = new EmbeddedGraphDatabase(data-dir);
lockManager =
I am looking at this in relation to an graph collections class,
UnrolledLinkedList. Since this is a library class that others might use it
has to rely on the top level interface, or more correctly it has to be able
to work off being passed a Node.
Since Node has getGraphDatabase() returning a
Hi Guys,
We managed to fix the performance issue.
* Used Paramterised Queries - increased performance by 1 extra node per second,
reduced frequency usage of heap thus no jvm memory errors
* Configured our .Net client to use 70 threads for rest calls (I chose this
number as Neo4j reserves 80,
How man nodes are in uarr in the follow case?
UserNode[] uarr =
Neo4jUserUtils.createUserNodeListFromTraverse2(travDesc.traverse(underlyingNode));
Can you describe what your pages represent and what data representation you
are creating in the code?
Without looking at your code too closely it
Thad, there is some good advice contained within those links, but it doesn't
really fit the situation that I am discussing. The locks I am talking about
are locks on data within the graph database, not java programmatic locks
(though in a non-HA environment java locks could well be used as an
Ah, transaction awareness, gotcha.
On Sep 25, 2011 10:19 PM, Bryce bryc...@gmail.com wrote:
Thad, there is some good advice contained within those links, but it
doesn't
really fit the situation that I am discussing. The locks I am talking
about
are locks on data within the graph database, not
OK Thank you. i'll try it. But get node's property is still a problem,When I
need more threads access one node's property.
The uarr just have 40 thousand fans. If I want finish the sort will cost
800ms once. This is notqualified . Now I changed my code. Got the property
before sort. Put into an
16 matches
Mail list logo