[
https://issues.apache.org/jira/browse/TRAFODION-14?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Selvaganesan Govindarajan resolved TRAFODION-14.
------------------------------------------------
Resolution: Fixed
Fix Version/s: 2.0-incubating
It is clear HBase needs to have the timestamp dimension in the rows to be more
granular than the milliseconds granularity. In case of index maintenance it was
possible for the Trafodion engine to avoid this problem by skipping the delete.
However, this issue can occur if the user application tries to delete and
insert the same row in succession within a milliseconds interval.
> upsert or merge into a table with indexes can result in inconsistency between
> index and table
> ---------------------------------------------------------------------------------------------
>
> Key: TRAFODION-14
> URL: https://issues.apache.org/jira/browse/TRAFODION-14
> Project: Apache Trafodion
> Issue Type: Bug
> Reporter: Selvaganesan Govindarajan
> Assignee: Selvaganesan Govindarajan
> Priority: Critical
> Fix For: 2.0-incubating
>
>
> Steps to reproduce this problem
> set schema seabase ;
> CREATE TABLE TRAFODION.SEABASE.TEST1
> (
> A INT NO DEFAULT NOT NULL NOT DROPPABLE
> , B INT DEFAULT NULL
> , PRIMARY KEY (A ASC)
> )
> ;
> CREATE INDEX IDX1 ON TRAFODION.SEABASE.TEST1
> (
> B ASC
> )
> ;
> CREATE TABLE TRAFODION.SEABASE.TEST2
> (
> A INT NO DEFAULT NOT NULL NOT DROPPABLE
> , B INT DEFAULT NULL
> , PRIMARY KEY (A ASC)
> )
> ;
> insert into test2 values
> (1,1),(2,2),(3,3),(4,4),(5,5),(6,6),(7,7),(8,8),(9,9),(10,10) ;
> prepare s2 from upsert into test1 select * from test2;
> 1) execute s2;
> 2) select * from test1 ;
> If you repeat steps 1 and 2 more than once, at times fewer than 10 rows are
> selected from the step 2.
> In this particular case select * from test1 involves index scan plan.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
