[ 
https://issues.apache.org/jira/browse/FLINK-19101?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17187647#comment-17187647
 ] 

fa zheng commented on FLINK-19101:
----------------------------------

Hi [~danny0405], you can create t1 and t2 freely as long as the fields are 
string type. The reason is obvious, you can see the comments in above code



> The SelectivityEstimator throw an NullPointerException when 
> convertValueInterval with string type
> -------------------------------------------------------------------------------------------------
>
>                 Key: FLINK-19101
>                 URL: https://issues.apache.org/jira/browse/FLINK-19101
>             Project: Flink
>          Issue Type: Bug
>          Components: Table SQL / Planner
>    Affects Versions: 1.10.2, 1.11.1
>            Reporter: fa zheng
>            Priority: Critical
>             Fix For: 1.12.0
>
>
> For a SQL : 
> {code:sql}
> select t1.a, t2.b from t1  join t2  on t1.b=t2.b and t1.b<'2'
> {code}
> It will throw java.lang.NullPointerException because SelectivityEstimator 
> convert character value interval  to string without considering null 
> situation.
> {code:scala}
>   def convertValueInterval(
>       interval: ValueInterval,
>       typeFamily: RelDataTypeFamily): ValueInterval = {
>     require(interval != null && typeFamily != null)
>     interval match {
>       case ValueInterval.empty | ValueInterval.infinite => interval
>       case _ =>
>         val (lower, includeLower) = interval match {
>           case li: WithLower => (li.lower, li.includeLower)
>           case _ => (null, false)
>         }
>         val (upper, includeUpper) = interval match {
>           case ui: WithUpper => (ui.upper, ui.includeUpper)
>           case _ => (null, false)
>         }
>         typeFamily match {
>           case SqlTypeFamily.NUMERIC | SqlTypeFamily.BOOLEAN | 
> SqlTypeFamily.DATE |
>                SqlTypeFamily.TIME | SqlTypeFamily.TIMESTAMP =>
>             ValueInterval(
>               comparableToDouble(lower),
>               comparableToDouble(upper),
>               includeLower,
>               includeUpper)
>           case SqlTypeFamily.CHARACTER =>
>             ValueInterval(
>               lower.toString,     //It can't call toString directly
>               upper.toString     //It can't call toString directly
>               includeLower,
>               includeUpper)
>           case _ => throw new UnsupportedOperationException(s"Unsupported 
> typeFamily: $typeFamily")
>         }
>     }
>   }
> {code}



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to