You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@s2graph.apache.org by "ASF GitHub Bot (JIRA)" <ji...@apache.org> on 2016/02/23 07:17:18 UTC
[jira] [Commented] (S2GRAPH-21) Change PostProcessBenchmarkSpec not
to store and fetch test data from storage
[ https://issues.apache.org/jira/browse/S2GRAPH-21?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15158363#comment-15158363 ]
ASF GitHub Bot commented on S2GRAPH-21:
---------------------------------------
Github user emeth-kim commented on the pull request:
https://github.com/apache/incubator-s2graph/pull/7#issuecomment-187560656
+1
> Change PostProcessBenchmarkSpec not to store and fetch test data from storage
> -----------------------------------------------------------------------------
>
> Key: S2GRAPH-21
> URL: https://issues.apache.org/jira/browse/S2GRAPH-21
> Project: S2Graph
> Issue Type: Test
> Reporter: DOYUNG YOON
> Assignee: DOYUNG YOON
> Priority: Trivial
> Labels: newbie, test
> Original Estimate: 48h
> Remaining Estimate: 48h
>
> What PostProcessBenchmarkSpec supposed to do is testing performance of methods on PostProcess.
> Currently, this spec does following on setup phase.
> # delete and re-create test Service
> # delete and re-create test Label
> # insert test data into HBase
> also to run PostProcess`s method, it actually fetch inserted data from setup phase.
> I suggest not create test Label and insert test data entirely.
> instead we can create QueryRequest, QueryResult, QueryRequestWithResult classes without actually store test data.
> I think by using mocked edges
> # we can run this test without need for DBMS and HBase so we can concentrate on PostProcess logic`s performance only.
> # more flexible and fast test can be possible with regard to number of edges in QueryResult.
> In general, I think there should be no dependency on RDBMS + HBase for benchmark specs.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)