[
https://issues.apache.org/jira/browse/HIVE-20639?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
slim bouguerra updated HIVE-20639:
----------------------------------
Description:
This patch adds multiple record writers to allow Hive user writing data
directly to a Kafka Topic.
The end goal is have a writer that provides multiple write semantics modes.
* A At_most_once where all the records will be delivered with no guarantee or
reties eg best effort mode. (Done)
* B At_least_once, each record will be delivered with retries from the Kafka
Producer and Hive Write Task. (Done)
* C Exactly_once , Writer will be using Kafka Transaction API to ensure that
each record is delivered once. (no so, still need more discussions)
In addition to the new feature i have refactored the existing code to make it
more readable.
was:
This patch adds multiple record writers to allow Hive user writing data
directly to a Kafka Topic.
The writer provides multiple write semantics modes.
* A None where all the records will be delivered with no guarantee or reties.
* B At_least_once, each record will be delivered with retries from the Kafka
Producer and Hive Write Task.
* C Exactly_once , Writer will be using Kafka Transaction API to ensure that
each record is delivered once.
In addition to the new feature i have refactored the existing code to make it
more readable.
> Add ability to Write Data from Hive Table/Query to Kafka Topic
> --------------------------------------------------------------
>
> Key: HIVE-20639
> URL: https://issues.apache.org/jira/browse/HIVE-20639
> Project: Hive
> Issue Type: New Feature
> Components: kafka integration
> Reporter: slim bouguerra
> Assignee: slim bouguerra
> Priority: Major
> Attachments: HIVE-20639.2.patch, HIVE-20639.patch
>
>
> This patch adds multiple record writers to allow Hive user writing data
> directly to a Kafka Topic.
> The end goal is have a writer that provides multiple write semantics modes.
> * A At_most_once where all the records will be delivered with no guarantee or
> reties eg best effort mode. (Done)
> * B At_least_once, each record will be delivered with retries from the Kafka
> Producer and Hive Write Task. (Done)
> * C Exactly_once , Writer will be using Kafka Transaction API to ensure that
> each record is delivered once. (no so, still need more discussions)
> In addition to the new feature i have refactored the existing code to make it
> more readable.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
