[
https://issues.apache.org/jira/browse/KYLIN-4457?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17090333#comment-17090333
]
wangrupeng edited comment on KYLIN-4457 at 4/23/20, 7:15 AM:
-------------------------------------------------------------
It's not a bug. This problem happened because that when build segment with
range(2011-01-01 00:00:00, 2013-01-01 00:00:00), kylin will not process the
data in the date of '2013-01-01'. And spark will include the data of
'2013-01-01'. We can verify correctness of cube by using the following sql.
select count * FROM TEST_KYLIN_FACT LEFT JOIN TEST_ORDER ON
TEST_KYLIN_FACT.ORDER_ID = TEST_ORDER.ORDER_ID where CAL_DT > cast('2011-01-01'
as date) and CAL_DT < cast('2013-01-01' as date)
was (Author: wangrupeng):
It's not a bug. This problem happened because that when build segment with
range(2011-01-01 00:00:00, 2013-01-01 00:00:00), kylin will not process the
data in the date of '2013-01-01'. And spark will include the data of
'2013-01-01'. We can verify correctness of cube by using the following sql.
select count(*) FROM TEST_KYLIN_FACT LEFT JOIN TEST_ORDER ON
TEST_KYLIN_FACT.ORDER_ID = TEST_ORDER.ORDER_ID where CAL_DT > cast('2011-01-01'
as date) and CAL_DT < cast('2013-01-01' as date)
> Query cube result doesn't math with spark sql
> ---------------------------------------------
>
> Key: KYLIN-4457
> URL: https://issues.apache.org/jira/browse/KYLIN-4457
> Project: Kylin
> Issue Type: Bug
> Components: Query Engine, Storage - Parquet
> Reporter: xuekaiqi
> Assignee: wangrupeng
> Priority: Major
> Fix For: v4.0.0-beta
>
> Original Estimate: 16h
> Remaining Estimate: 16h
>
> Source: local CSV
> Result:
> ||cube||spark||
> |5010|5030|
> SQL:
> {code:java}
> select count(*) FROM TEST_KYLIN_FACT LEFT JOIN TEST_ORDER ON
> TEST_KYLIN_FACT.ORDER_ID = TEST_ORDER.ORDER_ID where CAL_DT > '2011-01-01
> 00:00:00' and CAL_DT < '2013-01-01 00:00:00'
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
