[SPARK-9740] [SPARK-9592] [SPARK-9210] [SQL] Change the default behavior of First/Last to RESPECT NULLS.#8113
[SPARK-9740] [SPARK-9592] [SPARK-9210] [SQL] Change the default behavior of First/Last to RESPECT NULLS.#8113yhuai wants to merge 8 commits into
Conversation
|
@hvanhovell I am making the default in this PR. |
|
Test build #40531 has finished for PR 8113 at commit
|
|
One final question, shouldn't we introduce a |
There was a problem hiding this comment.
We could for now simplify this to Literal.create(true, BooleanType).
|
yeah. I think that only requires a small change. Let me add it. |
|
Besides the |
|
@hvanhovell Alright. I added the |
There was a problem hiding this comment.
@ggupta81 This is the fix of SPARK-9592 for our master and 1.5 branch.
There was a problem hiding this comment.
@yhuai Both if and else branches are executing the same code.
|
Should i close my original pull request now? On Wed, Aug 12, 2015 at 10:38 AM, UCB AMPLab notifications@github.com
_Gaurav Gupta_Engineering Manager @ Adobe |
|
Test build #40589 has finished for PR 8113 at commit
|
|
Test build #1488 has finished for PR 8113 at commit
|
There was a problem hiding this comment.
Nit: Or(valueSet, Not(IsNull(child))) is a bit shorter. It is a matter of preference though...
|
One more small thing. We should probably also add the |
There was a problem hiding this comment.
Nit: Or(valueSet.left. valueSet.right) is shorter.
|
@hvanhovell Since we already pass the feature freeze deadline, I will not add new interfaces to DataFrame API (this PR is mainly about fixing the default behavior). If users request adding DF functions that expose |
|
Test build #40705 has finished for PR 8113 at commit
|
|
Test build #40724 has finished for PR 8113 at commit
|
|
Test build #1542 has finished for PR 8113 at commit
|
|
Test build #1551 timed out for PR 8113 at commit |
…pression1. https://issues.apache.org/jira/browse/SPARK-9592 #8113 has the fundamental fix. But, if we want to minimize the number of changed lines, we can go with this one. Then, in 1.6, we merge #8113. Author: Yin Huai <yhuai@databricks.com> Closes #8172 from yhuai/lastFix and squashes the following commits: b28c42a [Yin Huai] Regression test. af87086 [Yin Huai] Fix last. (cherry picked from commit 772e7c1) Signed-off-by: Michael Armbrust <michael@databricks.com>
…pression1. https://issues.apache.org/jira/browse/SPARK-9592 #8113 has the fundamental fix. But, if we want to minimize the number of changed lines, we can go with this one. Then, in 1.6, we merge #8113. Author: Yin Huai <yhuai@databricks.com> Closes #8172 from yhuai/lastFix and squashes the following commits: b28c42a [Yin Huai] Regression test. af87086 [Yin Huai] Fix last.
Conflicts: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregate/functions.scala sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregates.scala
|
test this please |
|
Test build #43948 has finished for PR 8113 at commit
|
|
Thanks! Merging to master. |
I am changing the default behavior of
First/Lastto respect null values (the SQL standard default behavior).https://issues.apache.org/jira/browse/SPARK-9740