[jira] [Work logged] (HIVE-26366) PTF Vectorization: BoundaryScanner for boolean

[ASF GitHub Bot (Jira)]

     [ 
https://issues.apache.org/jira/browse/HIVE-26366?focusedWorklogId=790333&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-790333
 ]

ASF GitHub Bot logged work on HIVE-26366:
-----------------------------------------

                Author: ASF GitHub Bot
            Created on: 13/Jul/22 09:04
            Start Date: 13/Jul/22 09:04
    Worklog Time Spent: 10m 
      Work Description: ayushtkn commented on code in PR #3412:
URL: https://github.com/apache/hive/pull/3412#discussion_r919833948


##########
ql/src/test/queries/clientpositive/vector_ptf_bounded_start.q:
##########
@@ -726,4 +727,51 @@ EXPLAIN VECTORIZATION DETAIL select p_mfgr, p_name, 
p_timestamp, rowindex, p_dat
                                     count(*) over(partition by p_varchar) as 
cs,
                                     sum(p_retailprice) over(partition by 
p_varchar) as s
 from vector_ptf_part_simple_orc
-where p_varchar = 'DA';
\ No newline at end of file
+where p_varchar = 'DA';
+
+select "************ BOOLEAN WINDOW RANGE TYPE ************";
+set hive.vectorized.execution.ptf.enabled=false;
+
+select "************ BOOLEAN WINDOW RANGE TYPE (NON-VECTORIZED REFERENCE) 
************";
+select p_mfgr, p_name, p_timestamp, rowindex, p_date, p_retailprice, p_char, 
p_boolean,
+       count(*) over(partition by p_mfgr order by p_boolean range between 3 
preceding and current row) as cs,
+       sum(p_retailprice) over(partition by p_mfgr order by p_boolean range 
between 3 preceding and current row) as s
+from vector_ptf_part_simple_orc;
+
+select "************ BOOLEAN WINDOW RANGE TYPE (VECTORIZED) ************";
+set hive.vectorized.execution.ptf.enabled=true;
+
+EXPLAIN VECTORIZATION DETAIL select p_mfgr, p_name, p_timestamp, rowindex, 
p_date, p_retailprice,p_char, p_boolean,
+                                    count(*) over(partition by p_mfgr order by 
p_boolean range between 3 preceding and
+                                    current row) as cs,
+                                    sum(p_retailprice) over(partition by 
p_mfgr order by p_boolean range between 3 preceding and current row) as s
+from vector_ptf_part_simple_orc;
+
+select p_mfgr, p_name, p_timestamp, rowindex, p_date, p_retailprice,p_char, 
p_boolean,

Review Comment:
   yeps, it was something screwed at the distance calculation logic. I fixed 
it. Now it seems identical.
   <img width="1566" alt="image" 
src="https://user-images.githubusercontent.com/25608848/178695198-079c542f-abdc-45a0-8412-148867dd85c9.png";>
   





Issue Time Tracking
-------------------

    Worklog Id:     (was: 790333)
    Time Spent: 40m  (was: 0.5h)

> PTF Vectorization: BoundaryScanner for boolean
> ----------------------------------------------
>
>                 Key: HIVE-26366
>                 URL: https://issues.apache.org/jira/browse/HIVE-26366
>             Project: Hive
>          Issue Type: Bug
>            Reporter: László Bodor
>            Assignee: Ayush Saxena
>            Priority: Major
>              Labels: pull-request-available
>          Time Spent: 40m
>  Remaining Estimate: 0h
>




--
This message was sent by Atlassian Jira
(v8.20.10#820010)