Spark aggreation by partition could use metadata files

[lrpt]

Hello everybody,
I have a apache iceberg table in aws glue, this table is partitioned by string year-month.
When I do a spark.sql("select count(1),partition_field from table group by partition_field"). Spark goes through every file and perform the count. Cant spark engine use just the metadata files as each underlying data file just contains data from one partition.
(I dont have any delete file)

Thanks.

[RussellSpitzer]

Spark does have pushdowns for this. I don't know what Spark version you are using but aggregate pushdown is implemented by @huaxingao