[ 
https://issues.apache.org/jira/browse/CALCITE-5591?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17727573#comment-17727573
 ] 

Dan Zou commented on CALCITE-5591:
----------------------------------

[~tanclary] I have the germ of an idea: 
{code:java}
  public static final SqlReturnTypeInference FORCE_NULLABLE_SUM =
      new SqlReturnTypeInferenceChain(DECIMAL_SUM_NULLABLE, LEAST_RESTRICTIVE)
          .andThen(SqlTypeTransforms.FORCE_NULLABLE);
{code}

And I find it is more complicated to handle overflow for each return type, it 
would be great if you could share some ideas when you find a good way to do it.


> Implement BigQuery functions SAFE_ADD, SAFE_SUBTRACT, SAFE_MULTIPLY, 
> SAFE_DIVIDE and SAFE_NEGATE
> ------------------------------------------------------------------------------------------------
>
>                 Key: CALCITE-5591
>                 URL: https://issues.apache.org/jira/browse/CALCITE-5591
>             Project: Calcite
>          Issue Type: New Feature
>            Reporter: Dan Zou
>            Assignee: Dan Zou
>            Priority: Major
>
> Add support for SAFE_ADD, SAFE_SUBTRACT, SAFE_MULTIPLY, SAFE_DIVIDE and 
> SAFE_NEGATE in BigQuery.
> Function description:
> * SAFE_ADD: Equivalent to the addition operator '+', but returns NULL if 
> overflow occurs.
> * SAFE_SUBTRACT: Returns the result of Y subtracted from X. Equivalent to the 
> subtraction operator '-', but returns NULL if overflow occurs.
> * SAFE_MULTIPLY: Equivalent to the multiplication operator '*', but returns 
> NULL if overflow occurs.
> * SAFE_DIVIDE: Equivalent to the division operator '/ ', but returns NULL if 
> an error occurs, such as a division by zero error.
> * SAFE_NEGATE: Equivalent to the unary minus operator '-', but returns NULL 
> if overflow occurs.
> See more details in [BigQuery 
> Doc|https://cloud.google.com/bigquery/docs/reference/standard-sql/functions-and-operators#safe_divide]



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

Reply via email to