Attila Szabo commented on SQOOP-3029:

Hi [~markuskem...@me.com],

Yes I'm suggesting to introduce an option for that, because since we'd 
introduced the quoting for Oracle (to being able to support special characters, 
spaces, etc. in table and column names) it's a real issue/source of problem 
that we're looking for column names with different case then it's available in 
HCatalog shema for example (but would be able to match both of them in 
uppercase for example).

What do you think, does it make sense on your side too?


> Add an option for uppercase/lowercase column name mapping between HCatalog 
> and RDBMS cloumn name list
> -----------------------------------------------------------------------------------------------------
>                 Key: SQOOP-3029
>                 URL: https://issues.apache.org/jira/browse/SQOOP-3029
>             Project: Sqoop
>          Issue Type: Improvement
>            Reporter: Attila Szabo
>             Fix For: 1.4.7
> The current implementation of HCatalog related export works out of the box 
> with RDBMS systems (e.g. Oracle) if the column names match case sensitive, 
> although in several cases the column names are received as an UPPER_CASE list 
> from the RDBMS, and thus the export fails, within the ClassWriter.
> Although the users have the possibility to define the name of the columns 
> explicit with the --columns option, in case of 100+ columns it's not 
> practical.
> It would be great to have an option which could do the matching in lower and 
> UPPER case as well. The start point for this implementation should start from 
> SqoopHCatUtilities#configureHCat method with respect to ensure HCatalog api 
> is able to handle the different cased column names.

This message was sent by Atlassian JIRA

Reply via email to