flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "James Cao (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-1919) Add HCatOutputFormat for Tuple data types
Date Thu, 30 Jul 2015 14:51:05 GMT

    [ https://issues.apache.org/jira/browse/FLINK-1919?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14647727#comment-14647727

James Cao commented on FLINK-1919:


> Add HCatOutputFormat for Tuple data types
> -----------------------------------------
>                 Key: FLINK-1919
>                 URL: https://issues.apache.org/jira/browse/FLINK-1919
>             Project: Flink
>          Issue Type: New Feature
>          Components: Java API, Scala API
>            Reporter: Fabian Hueske
>            Assignee: James Cao
>            Priority: Minor
>              Labels: starter
> It would be good to have an OutputFormat that can write data to HCatalog tables.
> The Hadoop `HCatOutputFormat` expects `HCatRecord` objects and writes these to HCatalog
tables. We can do the same thing, by creating these `HCatRecord` object with a Map function
that precedes a `HadoopOutputFormat` that wraps the Hadoop `HCatOutputFormat`.
> Better support for Flink Tuples can be added by implementing a custom `HCatOutputFormat`
that also depends on the Hadoop `HCatOutputFormat` but internally converts Flink Tuples to
`HCatRecords`. This would also include to check if the schema of the HCatalog table and the
Flink tuples match. For data types other than tuples, the OutputFormat could either require
a preceding Map function that converts to `HCatRecords` or let users specify a MapFunction
and invoke that internally.
> We have already a Flink `HCatInputFormat` which does this in the reverse directions,
i.e., it emits Flink Tuples from HCatalog tables.

This message was sent by Atlassian JIRA

View raw message