Re: Import graph from ConfiguredGraphFactory to hbase


HadoopMarc <bi...@...>
 

First note that the BulkLoaderVertexProgram was deprecated in TinkerPop. You might want to see:
https://groups.google.com/g/janusgraph-users/c/d3F4dZdQ2XQ/m/HolppzItCAAJ

For working around your issue you can consult the API docs.

The getConfiguration() method gives you a configuration Map:
https://javadoc.io/doc/org.janusgraph/janusgraph-core/latest/org/janusgraph/core/ConfiguredGraphFactory.html

The BulkLoaderVertexProgram.builder class has a writeGraph(Configuration) method. Of course, you need some juggling to get the Map into the Configuration object.

HTH,   Marc

Op donderdag 22 oktober 2020 om 15:14:50 UTC+2 schreef mor...@...:


Hello everyone, we want to test some OLAP queries for our requests,and, as far,as I understand, to do that, graph needs to be exported to Hadoop.

* janusgraph-0.5.1, hadoop-2.10.1

Currently we access out graphs using configuredGraphFactory.open('graphName') and our graph uses configuration listed below:

    gremlin.graph=org.janusgraph.core.ConfiguredGraphFactory
    storage.backend=hbase
    graph.graphname=ConfigurationManagementGraph

Now, to load graph to Hadoop,I am trying  to do something like this (using defaut Hadoop/spark  load configuration file hadoop-load.properties):

<gremlin console>
:plugin use tinkerpop.spark
graph = GraphFactory.open('conf/hadoop-graph/hadoop-load.properties')
blvp = BulkLoaderVertexProgram.build().writeGraph(<need to choose specific graph from ConfiguredGraphFactory here>).create(graph)                graph.compute(SparkGraphComputer).program(blvp).submit().get() graph.close()

But I have no idea how to access specific graph from factory here, because writeGraph() takes configuration file as input

Join janusgraph-users@lists.lfaidata.foundation to automatically receive all group messages.