Reading Iceberg tables show 0.0B for the Input column on Spark's executor UI

[hsiang-c]

Describe the bug

spark.comet.scan.icebergNative.enabled=true

spark.comet.scan.icebergNative.enabled=false

Steps to reproduce

SELECT COUNT(DISTINCT(col_string)) FROM $ICEBERG_TABLE WHERE ... (predicates)

Expected behavior

CometIcebergNativeScan should also report input size per executor.

Additional context

No response

[hsiang-c]

take