[
https://issues.apache.org/jira/browse/HIVE-21742?focusedWorklogId=253397&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-253397
]
ASF GitHub Bot logged work on HIVE-21742:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 03/Jun/19 21:50
Start Date: 03/Jun/19 21:50
Worklog Time Spent: 10m
Work Description: vineetgarg02 commented on pull request #656: HIVE-21742
: Vectorization: CASE result type casting
URL: https://github.com/apache/hive/pull/656#discussion_r290052428
##########
File path: ql/src/test/results/clientpositive/perf/spark/query80.q.out
##########
@@ -782,7 +782,7 @@ STAGE PLANS:
1 Map 33
Statistics: Num rows: 231917759 Data size: 31534108438 Basic
stats: COMPLETE Column stats: NONE
Select Operator
- expressions: _col15 (type: string), _col5 (type:
decimal(7,2)), CASE WHEN (_col9 is not null) THEN (_col9) ELSE (0) END (type:
decimal(12,2)), (_col6 - CASE WHEN (_col10 is not null) THEN (_col10) ELSE (0)
END) (type: decimal(13,2))
+ expressions: _col15 (type: string), _col5 (type:
decimal(7,2)), CASE WHEN (_col9 is not null) THEN (_col9) ELSE (CAST( 0 AS
decimal(7,2))) END (type: decimal(7,2)), (_col6 - CASE WHEN (_col10 is not
null) THEN (_col10) ELSE (CAST( 0 AS decimal(7,2))) END) (type: decimal(8,2))
Review comment:
change to 7,2 is due to the `_col9`'s type which is decimal(7,2). The whole
expression's type is determined to be decimal(8,2) for some reason, but this
change is in line with previous logic where even though internal type was
decimal(12,2) whole case expr was decimal(13,2). In short type inference looks
to be more tight now.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 253397)
Time Spent: 2h 10m (was: 2h)
> Vectorization: CASE result type casting
> ---------------------------------------
>
> Key: HIVE-21742
> URL: https://issues.apache.org/jira/browse/HIVE-21742
> Project: Hive
> Issue Type: Bug
> Components: Logical Optimizer, Vectorization
> Affects Versions: 3.1.1
> Reporter: Gopal V
> Assignee: Vineet Garg
> Priority: Major
> Labels: pull-request-available
> Attachments: HIVE-21742.1.patch, HIVE-21742.2.patch,
> HIVE-21742.3.patch, HIVE-21742.4.patch
>
> Time Spent: 2h 10m
> Remaining Estimate: 0h
>
> {code}
> create temporary table foo(q548284 int);
> insert into foo values(1),(2),(3),(4),(5),(6);
> select q548284, CASE WHEN ((q548284 = 1)) THEN (0.2) WHEN ((q548284 = 2))
> THEN (0.4) WHEN ((q548284 = 3)) THEN (0.6) WHEN ((q548284 = 4)) THEN (0.8)
> WHEN ((q548284 = 5)) THEN (1) ELSE (null) END from foo order by q548284 limit
> 1;
> {code}
> Fails with
> {code}
> Caused by: java.lang.ClassCastException:
> org.apache.hadoop.hive.ql.exec.vector.LongColumnVector cannot be cast to
> org.apache.hadoop.hive.ql.exec.vector.DecimalColumnVector
> at
> org.apache.hadoop.hive.ql.exec.vector.DecimalColumnVector.setElement(DecimalColumnVector.java:130)
> at
> org.apache.hadoop.hive.ql.exec.vector.expressions.IfExprColumnNull.evaluate(IfExprColumnNull.java:101)
> {code}
> This gets fixed if the case return of (1) is turned into a (1.0).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
