S3 hive style partitioned writes

[arthurpassos]

Describe the new feature

A clear, concise description of the feature you'd like to see.

Use case

create table s3 (year uint16, class string, nick string) engine=s3(s3_conn, filename='{_partition_id}.parquet') partition by (year, class);

insert into s3 values (2022, 'warrior', 'aplayer');

use_hive should prolly be an engine setting
will add more details alter