Refactor the exit in a User-define procedure in Java

Drivers & Stacks
, ,
1

Hi everyone.
I'm working on a Java procedure.
I'm doing this:

@Procedure(name = "skynet.issue100028001", mode = Mode.READ)
    @Description("Rosetta Issue 000028-001")
    public Stream<MapResult> issue28return() {
        String query = "MATCH (t:partition)-[p:PARAMETER_PARTITION]->(e:param_file_parameter) " +
                "WITH t.ID_PK as idPartition,e.ID_PK as idPar, p.order as relac  " +
                "order by idPartition, relac asc  " +
                "WITH idPartition, collect([idPar,relac]) as list  " +
                "WITH idPartition as A , head(head(list)) as B  " +
                "RETURN  " +
                "toString(A) as AA,  " +
                "toString(B) as BB;";

        return tx.execute(query).stream().map(MapResult::new);

    }

when I call the procedure , the response is something like that

{
  "AA": "26227",
  "BB": "126110"
}
{
  "AA": "26229",
  "BB": "126092"
}
{
  "AA": "26243",
  "BB": "126051"
}

I'm trying to modify the output to get it by columns, but I can't get it.

AA              BB
"26227"    "126110"
"26229"    "126051"

Also taking into account that for other procedures the number of output fields can vary from one procedure to another.

I would appreciate any kind of help. Thanks in advance.

2

What does your MapResult object look like?

Just a comment, I am not sure there is benefit of writing a custom procedure just to execute a cypher query. I would think you would get the same performance executing the same query using a driver. The custom queries I wrote perform calculations I can't do in cypher, such as traversing graphs and calculating metrics, or complex updates. Have you don't any benchmarking to see if it provides benefit?