[FLINK-27247] ScalarOperatorGens.numericCasting is not compatible with legacy behavior

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 1.15.1, 1.16.0
• Component/s: Table SQL / Planner
• Labels: pull-request-available
• Assignee: xuyang
• Reporter: xuyang
• Created: 14/Apr/22 08:06
• Updated: 21/Apr/22 06:15
• Resolved: 21/Apr/22 06:15

Description

Add the following test cases in ScalarFunctionsTest:

// code placeholder
@Test
def test(): Unit ={
  testSqlApi("rand(1) + 1","")
} 

it will throw the following exception:

// code placeholder
org.apache.flink.table.planner.codegen.CodeGenException: Unsupported casting from DOUBLE to DOUBLE NOT NULL.
    at org.apache.flink.table.planner.codegen.calls.ScalarOperatorGens$.numericCasting(ScalarOperatorGens.scala:1734)
    at org.apache.flink.table.planner.codegen.calls.ScalarOperatorGens$.generateBinaryArithmeticOperator(ScalarOperatorGens.scala:85)
    at org.apache.flink.table.planner.codegen.ExprCodeGenerator.generateCallExpression(ExprCodeGenerator.scala:507)
    at org.apache.flink.table.planner.codegen.ExprCodeGenerator.visitCall(ExprCodeGenerator.scala:481)
    at org.apache.flink.table.planner.codegen.ExprCodeGenerator.visitCall(ExprCodeGenerator.scala:57)
    at org.apache.calcite.rex.RexCall.accept(RexCall.java:174)
    at org.apache.flink.table.planner.codegen.ExprCodeGenerator.$anonfun$visitCall$1(ExprCodeGenerator.scala:478)
    at scala.collection.TraversableLike.$anonfun$map$1(TraversableLike.scala:233)
    at scala.collection.mutable.ResizableArray.foreach(ResizableArray.scala:58)
    at scala.collection.mutable.ResizableArray.foreach$(ResizableArray.scala:51)
    at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:47)
    at scala.collection.TraversableLike.map(TraversableLike.scala:233)
    at scala.collection.TraversableLike.map$(TraversableLike.scala:226)
    at scala.collection.AbstractTraversable.map(Traversable.scala:104)
    at org.apache.flink.table.planner.codegen.ExprCodeGenerator.visitCall(ExprCodeGenerator.scala:469)
... 

This is because in ScalarOperatorGens#numericCasting,  FLINK-24779  lost the logic that in some cases there is no need to casting the left and right type.

Issue Links

• links to: GitHub Pull Request #19470