You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@carbondata.apache.org by "QiangCai (JIRA)" <ji...@apache.org> on 2017/11/08 07:28:00 UTC
[jira] [Resolved] (CARBONDATA-1582) 3. support StreamSinkProvider
and append batch data to row format file
[ https://issues.apache.org/jira/browse/CARBONDATA-1582?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
QiangCai resolved CARBONDATA-1582.
----------------------------------
Resolution: Fixed
> 3. support StreamSinkProvider and append batch data to row format file
> ----------------------------------------------------------------------
>
> Key: CARBONDATA-1582
> URL: https://issues.apache.org/jira/browse/CARBONDATA-1582
> Project: CarbonData
> Issue Type: Sub-task
> Components: data-load
> Reporter: QiangCai
> Assignee: QiangCai
>
> When data batch comes to carbon, carbon will check
> whether the batch id of the data is already committed. If yes, skip this batch,
> otherwise convert data fields and append to row format file directly. So, the
> records are either flushed or not flushed, data consistency is ensured.
> When writing to the row format file, carbon will convert the data fields the same way as for
> columnar format, like global dictionary conversion, timestamp conversion, complex type
> conversion, etc.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)