> On March 10, 2015, 5:58 p.m., Jarek Cecho wrote:
> > Overall looks good to me. I have one suggestion though - can we add test 
> > that will have two tables such that:
> > 
> > 1) First to import doesn't have PK and will use the new parameter to get to 
> > 1 mapper
> > 2) Second will have the PK and we will expect to use the full number of 
> > mappers
> > 
> > I'm thinking about test that will ensure that when running 
> > all-tables-import, the reset to single mapper will happen only for the 
> > affected table and not for other tables that happen to be listed after the 
> > affected one.

Good catch.  Since we are reusing the options that we provide to the import 
tool, the number of mappers will be reset and set to 1 for subsequent jobs.   I 
have cloned the options to pass to ImportTool.   One thing I am trying to do is 
to validate this from the tests - but that turned out to be lot harder than I 
thought.  I learnt today that mapred.map.tasks is ignored in local mode.   So 
adding a few test mechanics to handle that


> On March 10, 2015, 5:58 p.m., Jarek Cecho wrote:
> > src/docs/user/import.txt, line 219
> > <https://reviews.apache.org/r/31839/diff/1/?file=888590#file888590line219>
> >
> >     Nit: Trailing whitespace.

Will fix


On March 10, 2015, 5:58 p.m., Venkat Ranganathan wrote:
> > Jarcec

Thanks for the review.   The addition of new test turned out to be more work 
but worth it.   Will upload a new patch after validating


- Venkat


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/31839/#review75902
-----------------------------------------------------------


On March 9, 2015, 1:03 a.m., Venkat Ranganathan wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/31839/
> -----------------------------------------------------------
> 
> (Updated March 9, 2015, 1:03 a.m.)
> 
> 
> Review request for Sqoop.
> 
> 
> Repository: sqoop-trunk
> 
> 
> Description
> -------
> 
> Provides a new option autoreset-to-one-mapper which instructs sqoop to reset 
> the number of mappers to  one when primary key is not defined for a table and 
> no split by column is provided (which is the case for import-all-tables).   
> New tests added and documentation updated
> 
> 
> Diffs
> -----
> 
>   src/docs/user/import-all-tables-purpose.txt d5df88c 
>   src/docs/user/import-all-tables.txt 166825a 
>   src/docs/user/import.txt a6b23df 
>   src/java/org/apache/sqoop/SqoopOptions.java 0070d0b 
>   src/java/org/apache/sqoop/manager/SqlManager.java c81f170 
>   src/java/org/apache/sqoop/tool/BaseSqoopTool.java b41ee2d 
>   src/java/org/apache/sqoop/tool/ImportTool.java d5bf1eb 
>   src/test/com/cloudera/sqoop/TestAllTables.java 39e8941 
>   src/test/com/cloudera/sqoop/TestSqoopOptions.java 60460d9 
> 
> Diff: https://reviews.apache.org/r/31839/diff/
> 
> 
> Testing
> -------
> 
> All tests pass including newly added tests
> 
> 
> Thanks,
> 
> Venkat Ranganathan
> 
>

Reply via email to