Uploaded image for project: 'Flink'
  1. Flink
  2. FLINK-10019

singleton row type objects returned from sub query cannot be chained with other operators

Attach filesAttach ScreenshotVotersWatch issueWatchersCreate sub-taskLinkCloneUpdate Comment AuthorReplace String in CommentUpdate Comment VisibilityDelete Comments
    XMLWordPrintableJSON

    Details

      Description

      If explicitly return a CompositeType in udf.getResultType, will result in some failures in chained operators.
      For example: consider a simple UDF,

      object Func extends ScalarFunction {
        def eval(row: Row): Row = {
          row
        }
        override def getParameterTypes(signature: Array[Class[_]]): Array[TypeInformation[_]] =
          Array(Types.ROW(Types.INT))
        override def getResultType(signature: Array[Class[_]]): TypeInformation[_] =
          Types.ROW(Types.INT)
      }
      

      This should work perfectly since it's just a simple pass through, however

        @Test
        def testRowType(): Unit = {
          val data = List(
            Row.of(Row.of(12.asInstanceOf[Integer]), "1")
          )
          val env = StreamExecutionEnvironment.getExecutionEnvironment
          val stream = env.fromCollection(data)(Types.ROW(Types.ROW(Types.INT), Types.STRING))
      
          val tEnv = TableEnvironment.getTableEnvironment(env)
          val table = stream.toTable(tEnv, 'a, 'b)
          tEnv.registerFunction("func", Func)
          tEnv.registerTable("t", table)
      
          // This works perfectly
          val result1 = tEnv.sqlQuery("SELECT func(a) FROM t").toAppendStream[Row]
          result1.addSink(new StreamITCase.StringSink[Row])
      
          // This throws exception
          val result2 = tEnv.sqlQuery("SELECT func(a) as myRow FROM t").toAppendStream[Row]
          result2.addSink(new StreamITCase.StringSink[Row])
      
          env.execute()
        }
      

      Exception code:

      java.lang.IndexOutOfBoundsException: index (1) must be less than size (1)
      
      	at com.google.common.base.Preconditions.checkElementIndex(Preconditions.java:310)
      	at com.google.common.base.Preconditions.checkElementIndex(Preconditions.java:293)
      	at com.google.common.collect.SingletonImmutableList.get(SingletonImmutableList.java:41)
      	at org.apache.calcite.sql.type.InferTypes$2.inferOperandTypes(InferTypes.java:83)
      	at org.apache.calcite.sql.validate.SqlValidatorImpl.inferUnknownTypes(SqlValidatorImpl.java:1777)
      	at org.apache.calcite.sql.validate.SqlValidatorImpl.expandSelectItem(SqlValidatorImpl.java:459)
      	at org.apache.calcite.sql.validate.SqlValidatorImpl.expandStar(SqlValidatorImpl.java:349)
      ...
      

      This is due to the fact that Calcite inferOperandTypes does not expect to infer a struct RelDataType.

        Attachments

        Issue Links

          Activity

            People

            • Assignee:
              Unassigned
              Reporter:
              rongr Rong Rong

              Dates

              • Created:
                Updated:
                Resolved:

                Time Tracking

                Estimated:
                Original Estimate - Not Specified
                Not Specified
                Remaining:
                Remaining Estimate - 0h
                0h
                Logged:
                Time Spent - 20m
                20m

                  Issue deployment