[GitHub] [spark] cloud-fan commented on a change in pull request #31859: [SPARK-34769][SQL]AnsiTypeCoercion: return closest convertible type among TypeCollection

[GitBox <gi...@apache.org>]

cloud-fan commented on a change in pull request #31859:
URL: https://github.com/apache/spark/pull/31859#discussion_r596926825



##########
File path: sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/analysis/AnsiTypeCoercionSuite.scala
##########
@@ -377,10 +368,26 @@ class AnsiTypeCoercionSuite extends AnalysisTest {
       ArrayType(StringType, true),
       TypeCollection(ArrayType(StringType), StringType),
       ArrayType(StringType, true))
+
+    // When there are multiple convertible types in the `TypeCollection`, use the closet convertible
+    // data type among convertible types.
+    shouldCast(IntegerType, TypeCollection(BinaryType, FloatType, LongType), LongType)
+    shouldCast(IntegerType, TypeCollection(BinaryType, LongType, NumericType), IntegerType)
+    // If the result is Float type and Double type is also among the convertible target types,
+    // use Double Type instead of Float type.
+    shouldCast(LongType, TypeCollection(FloatType, DoubleType, StringType), DoubleType)
   }
 
   test("ineligible implicit type cast - TypeCollection") {
+    shouldNotCast(IntegerType, TypeCollection(StringType, BinaryType))
+    shouldNotCast(IntegerType, TypeCollection(BinaryType, StringType))
     shouldNotCast(IntegerType, TypeCollection(DateType, TimestampType))
+    shouldNotCast(IntegerType, TypeCollection(DecimalType(10, 2), StringType))
+    shouldNotCastStringInput(TypeCollection(NumericType, BinaryType))
+    // When there are multiple convertible types in the `TypeCollection` and there is no closest
+    // convertible data type among the convertible types.
+    shouldNotCastStringLiteral(TypeCollection(NumericType, BinaryType))

Review comment:
       This duplicates the previous case.




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org