Github user xccui commented on a diff in the pull request:
    --- Diff: 
    @@ -667,30 +719,62 @@ abstract class StreamTableEnvironment(
         // get CRow plan
         val plan: DataStream[CRow] = translateToCRow(logicalPlan, queryConfig)
    +    val rowtimeFields = logicalType
    +      .getFieldList.asScala
    +      .filter(f => FlinkTypeFactory.isRowtimeIndicatorType(f.getType))
    +    // convert the input type for the conversion mapper
    +    // the input will be changed in the OutputRowtimeProcessFunction later
    +    val convType = if (rowtimeFields.size > 1) {
    +      throw new TableException(
    --- End diff --
    I got an idea, but not sure if it's applicable. We allow multiple rowtime 
fields in a stream but only activate one in an operator. Since the timestamps 
are stored in records, the other inactive rowtime fields can just be taken as 
common fields. Any changes on the rowtime fields will render them invalid for 
rowtime use. IMO, there are not too many queries (maybe only over aggregate and 
join) depending on the rowtime, thus the optimizer may be able to deduce which 
rowtime field should be activated in an operator. However, some existing logics 
may be affected by that.

If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at or file a JIRA ticket
with INFRA.

Reply via email to