[ 
https://issues.apache.org/jira/browse/CALCITE-6379?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Mihai Budiu closed CALCITE-6379.
--------------------------------
    Resolution: Duplicate

Closing in favor of CALCITE-6685

> Arithmetic in Calcite does not produce results with the expected type
> ---------------------------------------------------------------------
>
>                 Key: CALCITE-6379
>                 URL: https://issues.apache.org/jira/browse/CALCITE-6379
>             Project: Calcite
>          Issue Type: Bug
>          Components: core
>    Affects Versions: 1.36.0
>            Reporter: Mihai Budiu
>            Priority: Major
>
> This seems to be a very old bug in Calcite, which appears in SqlOperatorTest 
> under the name FNL25. It has been around for at least 10 years, but it 
> affects other parts of the code as well. I could not find a more precise 
> description of this bug in the JIRA, so I am filing a new one. I think this 
> is at least a critical bug.
> The problem affects the expression evaluator in Calcite. After evaluating an 
> arithmetic operation the result should be cast to the expected result type. 
> For example, adding two SMALLINT values should cast the result to SMALLINT. 
> Currently the code generated by the RexToLixTranslator is missing this cast. 
> Since Java evaluation rules are different from SQL, many results are 
> incorrect.



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

Reply via email to