You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@sqoop.apache.org by "Veena Basavaraj (JIRA)" <ji...@apache.org> on 2014/10/04 00:57:37 UTC

[jira] [Created] (SQOOP-1566) Sqoop2: Fix the upgrade code with 1498 schema renames

Veena Basavaraj created SQOOP-1566:
--------------------------------------

             Summary: Sqoop2: Fix the upgrade code with 1498 schema renames
                 Key: SQOOP-1566
                 URL: https://issues.apache.org/jira/browse/SQOOP-1566
             Project: Sqoop
          Issue Type: Improvement
            Reporter: Veena Basavaraj
            Assignee: Veena Basavaraj


Gwen and I have this proposal.

Most importantly, not to have NULLs representing one too many things!!
 

SQ_CONNECTOR -> will be now call SQ_CONFIGURABLE ( it is basically a table to hold entities that will specify configs)

Add another field to it, that is CONFIGURABLE_TYPE ( it can be connector or driver at this point ). we can have a standard id for  Driver that we will use in the SQOOP, having nulls to represent things is not clean since "none/null" can mean different things at different contexts


Next,

We propose to split up the SQ_FORM ( which is the SQ_CONFIG after 1498) to SQ_JOB_CONFIG and SQ_LINK_CONFIG ( it will parallel the SQ_JOB_INPUT and SQ_LINK_INPUT)

Direction is only relevant to job config ( and henceforth the job inputs). So we will store the DIRECTION field in JOB_CONFIG.

If we want to have SQ_CONFIG, we can have it with the SQ_CFG_TYPE and then it will have  FK on the SQ_CONFIGURABLE, so we know which entity owns this config easily. 

We can then have SQ_JOB_CONFIG and SQ_LINK_CONFIG have a FK on SQ_CONFIG.

Alternatively, dont have SQ_CONFIG, and just have SQ_JOB_CONFIG and SQ_LINK_CONFIG and have a FK directly to SQ_CONFIGURABLE, both will have the CONFIG_NAME and INDEX fields.


Please vote




--
This message was sent by Atlassian JIRA
(v6.3.4#6332)