[HIVE-24666] Vectorized UDFToBoolean may unable to filter rows if input is string

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 4.0.0-alpha-1
• Component/s: Vectorization
• Labels: pull-request-available
• Assignee: Zhihua Deng
• Reporter: Zhihua Deng
• Created: 20/Jan/21 11:53
• Updated: 02/Oct/23 13:27
• Resolved: 16/Feb/21 22:29

Description

If we use cast boolean to filter rows,  in vectorization mode the filter may be unable to filter rows,  step to reproduce:

create table vtb (key string, value string);
insert into table vtb values('0', 'val0'), ('false', 'valfalse'),('off', 'valoff'),('no','valno'),('vk', 'valvk');
select distinct value from vtb where cast(key as boolean); 

It's seems we don't generate a SelectColumnIsTrue to filter the rows if the casted type is string:
 https://github.com/apache/hive/blob/ff6f3565e50148b7bcfbcf19b970379f2bd59290/ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorizationContext.java#L2995-L2996

Issue Links

• causes: HIVE-27754 Query Filter with OR condition updates every record in the table
• links to: GitHub Pull Request #1890