[
https://issues.apache.org/jira/browse/KYLIN-4839?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
xue lin updated KYLIN-4839:
---------------------------
Description:
Cube does not use the latest snapshot when build， and this will introduce
accuracy problems about data.
detailed case1:
the 2020-11-25 snapshot was used by cube between 2020-11-24 and 2020-11-25，but
when builded after 2020-11-25, it used the 2020-11-07 snapshot rather than the
latest snapshot , see fig1 fig2 fig3
detailed case2:
data source from hive changed on 20201201，the code trigger on 20201202. see fig4
so there should have 2 snapshots, one is before changed from hive ,the other is
after. see fig5
from fig6 we can see that the cubes used the 2020-12-01 snapshot between
20201126 and 20201201; between 20201201 and 20201202, between 20201204 and
20201205; between 20201205 and 20201206; between 20201206and 20201207;
see fig6
but the cubes used the 2020-12-02 snapshot between 20201202and 20201203;
between 20201203 and 20201204; see fig 7
was:
Cube does not use the latest snapshot when build， and this will introduce
accuracy problems about data.
detailed case1:
the 2020-11-25 snapshot was used by cube between 2020-11-24 and 2020-11-25，but
when builded after 2020-11-25, it used the 2020-11-07 snapshot rather than the
latest snapshot , see fig1 fig2 fig3
> Cube does not use the latest snapshot when build
> ------------------------------------------------
>
> Key: KYLIN-4839
> URL: https://issues.apache.org/jira/browse/KYLIN-4839
> Project: Kylin
> Issue Type: Bug
> Affects Versions: v3.0.2
> Reporter: xue lin
> Priority: Major
> Attachments: fig1.png, fig2.png, fig3.png, fig4.png
>
>
> Cube does not use the latest snapshot when build， and this will introduce
> accuracy problems about data.
> detailed case1:
> the 2020-11-25 snapshot was used by cube between 2020-11-24 and
> 2020-11-25，but when builded after 2020-11-25, it used the 2020-11-07 snapshot
> rather than the latest snapshot , see fig1 fig2 fig3
>
> detailed case2:
> data source from hive changed on 20201201，the code trigger on 20201202. see
> fig4
> so there should have 2 snapshots, one is before changed from hive ,the other
> is after. see fig5
> from fig6 we can see that the cubes used the 2020-12-01 snapshot between
> 20201126 and 20201201; between 20201201 and 20201202, between 20201204 and
> 20201205; between 20201205 and 20201206; between 20201206and 20201207;
> see fig6
> but the cubes used the 2020-12-02 snapshot between 20201202and 20201203;
> between 20201203 and 20201204; see fig 7
>
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
