[native] Iceberg basic insertion

[PingLiuPing]

Use this issue to track the problems during design and implement native basic iceberg insertion

Expected Behavior or Use Case

Presto Component, Service, or Connector

Possible Implementation

Example Screenshots (if appropriate):

Context

[PingLiuPing]

Velox PR: facebookincubator/velox#10996
Presto PR: #25389

[PingLiuPing]

1. In this PR, it only support identity transform. But decimal and varbinary are not supported as identity transform column. Since the partition name for varbinary is not correct. And for decimal, during reading there is error.

[PingLiuPing]

2. nested column type as partition column is not supported.

[PingLiuPing]

3. iceberg spec requires store data files under path_to/data/. The data part needs to be added by upper layer.
4. Velox can handle the ending / in target write path. Means with or without /, both are valid input to velox.

[PingLiuPing]

5. Velox iceberg writer only support parquet format now. ORC is not supported.

[PingLiuPing]

6. When integrating with Presto, the spec ID in commit metadata json is "partitionSpecJson".

[PingLiuPing]

7. An outstanding issue is when there are special chars in the column name, query result is wrong. This is found during integration with Gluten. From Presto, I don't see such issue.

[PingLiuPing]

8. When writing parquet data file, the iceberg field_id is not been wrote to parquet file. See https://iceberg.apache.org/spec/#parquet. This needs to be fixed