andygrove commented on code in PR #1667:
URL: https://github.com/apache/datafusion-comet/pull/1667#discussion_r2054613906
##########
spark/src/main/scala/org/apache/comet/CometSparkSessionExtensions.scala:
##########
@@ -161,10 +161,28 @@ class CometSparkSessionExtensions
}
if (!schemaSupported) {
- fallbackReasons += s"Unsupported schema ${scanExec.requiredSchema}
for $scanImpl"
+ val field: Option[StructField] =
+ CometNativeScanExec
+ .findUnsupportedField(scanExec.requiredSchema)
+
.orElse(CometScanExec.findUnsupportedField(scanExec.requiredSchema))
+ assert(field.nonEmpty)
+ fallbackReasons +=
+ s"Unsupported field ${field.get} in schema
${scanExec.requiredSchema} for $scanImpl"
Review Comment:
This PR introduces a two-step process:
1) call `isTypeSupported`
2) call `findUnsupportedField` if `isTypeSupported` returns `false`
I wonder if we should consider a different approach so that do not have to
duplicate the logic.
One idea would be to pass an accumulator into `isTypeSupported` that can
collect fallback reasons. For example:
```scala
def isTypeSupported(dt: DataType, fallbackReasons: mutable.List[String])
```
Another approach would be to rename `isTypeSupported` to
`validateTypeSupported` and have it throw an exception when it encounters the
first unsupported type.
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
