[
https://issues.apache.org/jira/browse/CARBONDATA-4043?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Indhumathi Muthumurugesh updated CARBONDATA-4043:
-------------------------------------------------
Description:
h3. When dimension is added in older versions like 1.1, by default it will be
sort column. In sort step we assume data will be coming as sort column in the
beginning. But the added column will be at last eventhough sort column. So,
while building the dataload configurations for loading data, we rearrange the
columns(dimensions and datafields) in order to bring the sort column to
beginning and no-sort to last and revert them back to schema order before
FinalMerge/DataWriter step.
Issue:
Data loading is failing because of castException in data writing step in case
of NO_SORT and in final sort step in case of LOCAL_SORT.
> Fix data load failure issue for columns added in legacy store
> -------------------------------------------------------------
>
> Key: CARBONDATA-4043
> URL: https://issues.apache.org/jira/browse/CARBONDATA-4043
> Project: CarbonData
> Issue Type: Bug
> Affects Versions: 2.1.0
> Reporter: Indhumathi Muthumurugesh
> Priority: Major
>
> h3. When dimension is added in older versions like 1.1, by default it will be
> sort column. In sort step we assume data will be coming as sort column in the
> beginning. But the added column will be at last eventhough sort column. So,
> while building the dataload configurations for loading data, we rearrange the
> columns(dimensions and datafields) in order to bring the sort column to
> beginning and no-sort to last and revert them back to schema order before
> FinalMerge/DataWriter step.
> Issue:
> Data loading is failing because of castException in data writing step in
> case of NO_SORT and in final sort step in case of LOCAL_SORT.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
