[
https://issues.apache.org/jira/browse/SQOOP-1872?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Veena Basavaraj updated SQOOP-1872:
-----------------------------------
Description:
The JODA time jars are added in the JobManager today because the IDF
implementation - CSVIntermediateDataFormat needs it.
this cannot scale if more like AvroIDF or JSON IDF is added. so expose a method
to set dependent jars in IDF API
was:
Currently the IDF API is in connector SDK package,
The right place for IDF API is sqoop-spi package since custom implementations
of the IDF can be done in future.
Instead today sqoop SPI that holds connectors depends on SDK which in fact
should be the other way round, the connector API is the root and should not
depend on its implemenation package.
connector SDK was intended to hold concrete implementations and some extent
utility classes that the connectors use, there is obvioulsy some leaks and
execution engine even depends on connecto-rsdk
> IDF API should expose a method to add dependent jars
> ----------------------------------------------------
>
> Key: SQOOP-1872
> URL: https://issues.apache.org/jira/browse/SQOOP-1872
> Project: Sqoop
> Issue Type: Bug
> Components: sqoop2-framework
> Reporter: Veena Basavaraj
> Assignee: Veena Basavaraj
> Labels: newbie
> Fix For: 1.99.5
>
>
> The JODA time jars are added in the JobManager today because the IDF
> implementation - CSVIntermediateDataFormat needs it.
> this cannot scale if more like AvroIDF or JSON IDF is added. so expose a
> method to set dependent jars in IDF API
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
