Fix failure when hidden partition column name conflicted in Iceberg

[ebyhr]

Description

Iceberg connector caused query failure when the hidden partition column name conflicts with other columns.

The approach is different from Spark. They throw an exception when conflicts happen:

CREATE TABLE test (ts timestamp, ts_day int) USING iceberg PARTITIONED BY (day(ts));
Failed in [create table test (ts timestamp, ts_day int) using iceberg partitioned by (day(ts))]
java.lang.IllegalArgumentException: Cannot create partition from name that exists in schema: ts_day

Fixes #22351

Release notes

(x) Release notes are required, with the following suggested text:

# Iceberg
* Fix failure when hidden partition name conflicted. ({issue}`22351`)

[findepi]

I am not against supporting explicit aliases, but i also think that auto-generated names shouldn't conflict with user-provided names without extra effort from the user, so I am not convinced this PR is needed.

IMO the logic in parsePartitionField should automatically deconflict the names.

Once we have this, the only remaning source of name collisions is when user wants to add a new column that overlaps a partitioning field. If we wanted to solve this problem too, the right resolution for this would be allow user to rename existing partitioning field.

[findinpath]

Did you by chance open a ticket on iceberg to have a more structured way of identifying such a problem?