[
https://issues.apache.org/jira/browse/IOTDB-1698?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17709600#comment-17709600
]
xiaozhihong commented on IOTDB-1698:
------------------------------------
If the large data volume performs correctness point-to-point verification, the
current change is to use two databases to compare directly.
> Record data set, data set query verification execution efficiency is slow
> -------------------------------------------------------------------------
>
> Key: IOTDB-1698
> URL: https://issues.apache.org/jira/browse/IOTDB-1698
> Project: Apache IoTDB
> Issue Type: Improvement
> Components: Benchmark
> Reporter: xiaozhihong
> Assignee: Hongyin Zhang
> Priority: Minor
>
> Execute steps:Take the record data set (locally generated TEXT file) as an
> example
> Step1:
> Generate a certain amount of data using the function of generating data sets
> (using out-of-order mode): it will take about half an hour to complete.
> The configuration is as follows:
> sg=20
> device=200
> sensor=300
> LOOP=8640
> BATCH_SIZE_PER_WRITE=100
> Use out-of-order mode 1, out-of-order 50%
> The data points written by each sensor under each device are: 864000
> Step2:
> Write the data set generated in the first step to IoTDB: it will take about
> 1.5 hours to complete.
> Step3:
> Use the query mode to verify whether the data in IoTDB is consistent with the
> records in the local TEXT.
> Start query time: 2021-09-14 17:26:18,693
> Deadline: 2021-09-15 16:21:01,378-nearly 23 hours of execution, and the
> progress is only 0.01%
> Issue:
> It is estimated that it will take nearly 10,000 days to complete the query,
> which is time-consuming and serious
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
