Apache TinkerPopAT
Apache TinkerPop•13mo ago
blacklight

mergeV with onMerge when extra properties are unknown

I'm in the following situation:

jobId = "spark:bdx_job_1"

// Initial vertices
vertices = [
    [
        (T.id): "spark:bdx_job_1",
        (T.label): "job",
        "url": single("spark://some/url"),
        "status": single("running"),
    ],
]

// First insertion
g.inject(vertices) \
    .unfold() \
    .mergeV() \
    .iterate()

// Updated vertices
updatedVertices = vertices = [
    [
        (T.id): jobId,
        (T.label): "job",
        "url": "spark://another/url",
        "status": "completed"
        }
    ]
]


I want to inject updatedVertices in such a way that the node is created if it doesn't exist, and its properties (excluding 
id
 and label) are updated if a match is found on <id, label>.

I tried this approach, but I'm in a situation where the extra properties are not known upfront, so the range/tail approach probably isn't feasible. I've tried ti tinker with sideEffect, but any attempts resulted in a serialization error through the JDK proxy. I've also tried some of the solutions suggested here, but I haven't had much luck either.
Stack Overflow
Efficiently bulk upserting vertices and edges containing properties...
Context
We're ingesting a combined total of between (500, 1500, 5000) (P25, P50, P75) vertices and edges per minute. Each vertex has roughly 5 "records" on it (id, label, 3 string propert...
Was this page helpful?