Github user EntilZha commented on a diff in the pull request:
https://github.com/apache/spark/pull/7580#discussion_r35295129
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/collectionOperations.scala
---
@@ -35,3 +36,47 @@ case class Size(child: Expression) extends
UnaryExpression with ExpectsInputType
nullSafeCodeGen(ctx, ev, c => s"${ev.primitive} = ($c).size();")
}
}
+
+case class ArrayContains(left: Expression, right: Expression) extends
BinaryExpression with ExpectsInputTypes {
+ override def dataType: DataType = BooleanType
+
+ override def inputTypes: Seq[AbstractDataType] = Seq(ArrayType,
AnyDataType)
--- End diff --
Agreed, we should not do implicit casting. The above python code should
work, but it is failing if I do an exact type match. Specifically, this code
makes it fail:
```scala
override def checkInputDataTypes(): TypeCheckResult = {
if (!left.dataType.isInstanceOf[ArrayType]) {
TypeCheckResult.TypeCheckFailure(
s"type of first input must be an array, not
${left.dataType.simpleString}")
} else {
val elementType = left.dataType.asInstanceOf[ArrayType].elementType
if (elementType.acceptsType(right.dataType) &&
right.dataType.acceptsType(elementType)) {
TypeCheckResult.TypeCheckSuccess
} else {
TypeCheckResult.TypeCheckFailure(
s"type of value must match array type " +
s"${elementType.simpleString}, not " +
s"${right.dataType.simpleString}")
}
}
}
```
due to this exception being thrown: `cannot resolve
'array_contains(data,1)' due to data type mismatch: type of value must match
array type bigint, not int;`
That seems like the wrong thing to do. Additionally, I think it should work
to do something like `array_contains([1.0, 2.0], 1)` and expect it to return
true. This can be left to `scala` I think when it does a comparison. This would
mean however that checking for exact type equality would break this. So either
we could leave this and allow comparison of obviously incompatible types, or we
could implement that the types of elements must match to some degree (for
example, numerics with numerics, strings with strings etc).
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
