Hi All,
Any advice on this.
Thanks
sam
On Sun, Nov 1, 2020 at 11:05 PM Anamika Solr <anamikasolr8...@gmail.com>
wrote:
> Hi All,
>
> I need to combine 3 different documents using hashjoin. I am using below
> query(ignore placeholder queries):
>
>
> hashJoin(hashJoin(search(collectionName,q="*:*",fl="id",qt="/export",sort="id
> desc"), hashed =
> select(search(collectionName,q="*:*",fl="id",qt="/export",sort="id
> asc")),on="id"), hashed =
> select(search(collectionName,q="*:*",fl="id",qt="/export",sort="id
> asc")),on="id")
>
> This works with simple TupleStream in java. But I also need to pass auth
> token on zk. So I have to use below code:
> ZkClientClusterStateProvider zkCluster = new
> ZkClientClusterStateProvider(zkHosts, null);
> SolrZkClient zkServer = zkCluster.getZkStateReader().getZkClient();
> StreamFactory streamFactory = new
> StreamFactory().withCollectionZkHost("collectionName"),
> zkServer.getZkServerAddress())
> .withFunctionName("search", CloudSolrStream.class)
> .withFunctionName("hashJoin", HashJoinStream.class)
> .withFunctionName("select", SelectStream.class);
>
> try (HashJoinStream hashJoinStream =
> (HashJoinStream)streamFactory.constructStream(expr);){}
>
> Issue is one hashjoin with nested select and search works fine with this
> api. But the multiple hashjoin is not completing the task. I can see
> expression is correctly parsed, but its waiting indefinitely to complete
> the thread.
>
> Any help is appreciated.
>
> Thanks,
> Anamika
>
