1 Answer
- Newest
- Most votes
- Most comments
1
Hello and thanks for the question. The ApacheTinkerPop Gremlin sample step currently uses an implementation that reads data into memory before taking a sample from it. That may change in the future, but for now are you able to put a limit in front of the step, so perhaps sample from something like limit(50000) ?
Depending upon the instance size being used, a larger instance type may sufficiently increase the memory enough for the whole sample to complete, but perhaps try with different limit sizes as a short term way to get at least some results.
So g.V().limit(50000).sample(1)
answered 13 days ago
Relevant content
- asked 2 months ago
- asked 6 months ago
AWS OFFICIALUpdated 2 years ago- AWS OFFICIALUpdated 7 months ago
- AWS OFFICIALUpdated 4 years ago
- AWS OFFICIALUpdated 4 months ago
Would there be a way to ensure I can tend towards capturing all nodes? If I wanted the random walk to occur across all nodes, would the limit impact the directions the walk could take?
For reference as well - this is a single instance DB, running now on r5.2xlarge
If you had something like
g.V().limit(100).sample(1).out()then yes it would mean the random walk could only ever start from one of the 100 found using thelimitstep.You could also potentially look at something like the
coinstep with a very small value, say perhapscoin(0.0001).sample(1)Gotcha - good to know, thank you! Might just have to run neptune exporter and not sample