Consolidate iceberg native-mode into the catalog type

[ChunxuTang]

As when the iceberg catalog type is HADOOP, the connector won't depend on the Hive metastore anymore, aka, in the native mode, we can safely remove the native mode from the iceberg connector config and only use the catalog type to determine whether it's in the native mode.

If the catalog type is HIVE, we use Hive metastore and it's in the non-native mode.
If the catalog type is HADOOP, we use HDFS to store both data and metadata and it's in the native mode.

Test plan - Unit test and integration test

== RELEASE NOTE ==

Iceberg Changes
* Remove the native-mode iceberg config. Use catalog.type instead.

[ChunxuTang]

Hi @beinan and @zhenxiao, could you have a review of this PR? Thanks!

[beinan]

nit: using switch? your call

[ChunxuTang]

Nice catch! I refactored the if/else if statements to switch.

[beinan]

might be out of scope of this PR, I'm thinking make this getIcebergTable logic to be a utility method or in polymorphism

[ChunxuTang]

Agree! I created a utility function to get the iceberg table.
Actually there are some other places such as IcebergMetadata and IcebergNativeMetadata where we could do some optimization by creating new abstraction layers. As this PR just focuses on merging the native-mode config into the catalog type, I'll send some follow-up PRs for these new abstraction layers.

[zhenxiao]

looks good @ChunxuTang could you please add release notes? This PR changes the Iceberg config, removes iceberg.native-mode, let's make corresponding doc changes

[ChunxuTang]

@zhenxiao Thanks for the review!
I updated the release note accordingly. Regarding the doc, I noticed that the doc is not updated since the support of iceberg native catalogs (#16612). I'll send a separate PR to update the doc.