Uh oh!
There was an error while loading. Please reload this page.
[SPARK-12629] [SparkR] Fixes for DataFrame saveAsTable method - #10580
[SPARK-12629] [SparkR] Fixes for DataFrame saveAsTable method#10580NarineK wants to merge 9 commits into
Conversation
SparkQA
commented
Jan 4, 2016
Test build #48692 has finished for PR 10580 at commit
|
SparkQA
commented
Jan 5, 2016
Test build #48697 has finished for PR 10580 at commit
|
felixcheung
commented
Jan 5, 2016
I think the direction we are heading is to have multiple sqlContext sessions and be able to specify which one to use, instead of automatically picking one? |
There was a problem hiding this comment.
I think the same problem exists in write.df()? Could you extract the following logic into a util function, and update write.df() also to use it?
There was a problem hiding this comment.
@felixcheung , @sun-rui , @shivaram : that's a good idea, I can use "getSqlContext()" method proposed in https://github.com/apache/spark/pull/9192/files#diff-b11442485f6b77bf47b58b4747321638R36
I guess I need to wait that pull request to go in and then I can merge it with mine .
There was a problem hiding this comment.
@felixcheung, it seems that you have it for saveAsTable too. After merge I can add my test cases ...
sun-rui
commented
Jan 5, 2016
@felixcheung, as we discussed before, I don't think it makes sense to support multiple sessions in SparkR now. Let's keep the current way until there is need to support it. |
There was a problem hiding this comment.
Please use df.write.saveAsTable() instead of deprecated df.saveAsTable().
shivaram
commented
Jan 6, 2016
NarineK
commented
Jan 6, 2016
sun-rui
commented
Jan 7, 2016
I think we may need more discussion on #9192, so let's merge this first. |
SparkQA
commented
Jan 7, 2016
Test build #48951 has finished for PR 10580 at commit
|
SparkQA
commented
Jan 7, 2016
Test build #48952 has finished for PR 10580 at commit
|
NarineK
commented
Jan 7, 2016
I see. I've pushed all my changes and left the piece of code which retrieves the context in case source is not passed by the argument. Later, you can replace it with the getContext utility method. |
shivaram
commented
Jan 19, 2016
felixcheung
commented
Jan 20, 2016
looks fine to me |
sun-rui
commented
Jan 20, 2016
@NarineK, could you update the write.df() function as well? |
NarineK
commented
Jan 21, 2016
updated write.df and saveDF |
SparkQA
commented
Jan 21, 2016
Test build #49840 has finished for PR 10580 at commit
|
SparkQA
commented
Jan 21, 2016
Test build #49884 has finished for PR 10580 at commit
|
NarineK
commented
Jan 21, 2016
added ! Let me know what do you think! |
sun-rui
commented
Jan 22, 2016
LGTM |
shivaram
commented
Jan 22, 2016
Merging this to master. |
shivaram
commented
Feb 2, 2016
@sun-rui@felixcheung I'm going to backport this to branch-1.6 so that the bug fix is available in 1.6.1. Let me know if you think there is a problem with it |
I've tried to solve some of the issues mentioned in: https://issues.apache.org/jira/browse/SPARK-12629 Please, let me know what do you think. Thanks! Author: Narine Kokhlikyan <narine.kokhlikyan@gmail.com> Closes#10580 from NarineK/sparkrSavaAsRable. (cherry picked from commit 8a88e12) Signed-off-by: Shivaram Venkataraman <shivaram@cs.berkeley.edu>
I've tried to solve some of the issues mentioned in: https://issues.apache.org/jira/browse/SPARK-12629
Please, let me know what do you think.
Thanks!