fix: map parquet field_id correctly (native_iceberg_compat)

Author: parthchandra

common/src/main/java/org/apache/comet/parquet/NativeBatchReader.java

@@ -283,6 +283,8 @@ public void init() throws Throwable {
                   sparkSchema.size(), requestedSchema.getFieldCount()));
         }
       }
+      // Convert clipped schema back to a Spark schema
+      sparkSchema = converter.convert(requestedSchema);
       this.parquetColumn =
           converter.convertParquetColumn(requestedSchema, Option.apply(this.sparkSchema));