-
Notifications
You must be signed in to change notification settings - Fork 28.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[SPARK-4508] [SQL] build native date type to conform behavior to Hive #3381
Conversation
Test build #23664 has started for PR 3381 at commit
|
Test build #23664 has finished for PR 3381 at commit
|
Test FAILed. |
retest this please. |
This is a weird build error. @shaneknapp |
Test build #23665 has started for PR 3381 at commit
|
Test build #23665 has finished for PR 3381 at commit
|
Test FAILed. |
looks like a compilation error, and the reason why no unit test results On Wed, Nov 19, 2014 at 11:56 PM, Daoyuan Wang notifications@github.com
|
Jenkins, test this please |
Test build #23676 has started for PR 3381 at commit
|
Test build #23676 has finished for PR 3381 at commit
|
Test FAILed. |
This builds successfully locally, and the build error is very confusing, since I never changed anything related to that. |
@marmbrus @rxin @liancheng Can you help verify this build error? I cloned this branch separately and the build is successful. |
The history of this build has already timed out... |
retest this please |
Test build #23816 has started for PR 3381 at commit
|
Test build #23816 has finished for PR 3381 at commit
|
Test FAILed. |
@liancheng can you reproduce the build error locally? I think it is a bug of Jenkins. |
i'll have the time to take a closer look at this tomorrow. On Mon, Nov 24, 2014 at 9:59 PM, Daoyuan Wang notifications@github.com
|
@adrian-wang Actually I can reproduce this compilation error with this: ./sbt/sbt -Pyarn -Phadoop-2.3 -Dhadoop.version=2.3.0 -Pkinesis-asl -Phive -Phive-0.12.0 -Phive-thriftserver clean compile Those profiles used by Jenkins are unfortunately not printed when compiling against Hive 0.12.0. You may refer to |
And the compilation failure is caused by this line https://github.com/apache/spark/pull/3381/files#diff-ff50aea397a607b79df9bec6f2a841dbL23 |
@liancheng thanks for your help! My local compilation with -Phive-0.13.1 is totally OK...... |
Test build #23827 has started for PR 3381 at commit
|
Test build #23827 has finished for PR 3381 at commit
|
Test PASSed. |
@adrian-wang Yea, Jenkins compiles Spark SQL with both Hive 0.12.0 and 0.13.1, and then runs SQL tests against 0.13.1. |
Hi @adrian-wang, thanks for working on this and I'm very sorry for letting it fall out of date. In general this seems like a reasonable change to make, but we do need to be very careful that we don't change the output types as those are part of the public API. I could be wrong, but it doesn't seem like we are doing that now. So I propose the following:
What do you think? |
Thanks for comments! Sorry for letting it out of date. I have no authority to reopen PRs, so I'll start a new session for this. |
Oh, sorry I thought you could reopen your own PRs. I guess that is not the case. Either way, please open a new one when ready. |
Store daysSinceEpoch as an Int value(4 bytes) to represent DateType, instead of using java.sql.Date(8 bytes as Long) in catalyst row. This ensures the same comparison behavior of Hive and Catalyst. Subsumes #3381 I thinks there are already some tests in JavaSQLSuite, and for python it will not affect python's datetime class. Author: Daoyuan Wang <daoyuan.wang@intel.com> Closes #3732 from adrian-wang/datenative and squashes the following commits: 0ed0fdc [Daoyuan Wang] fix test data a2fdd4e [Daoyuan Wang] getDate c37832b [Daoyuan Wang] row to catalyst f0005b1 [Daoyuan Wang] add date in sql parser and java type conversion 024c9a6 [Daoyuan Wang] clean some import order d6715fc [Daoyuan Wang] refactoring Date as Primitive Int internally 374abd5 [Daoyuan Wang] spark native date type support
Store daysSinceEpoch as an Int value(4 bytes) to represent DateType, instead of using java.sql.Date(8 bytes as Long) in catalyst row. This ensures the same comparison behavior of Hive and Catalyst.