HIVE-27234: Iceberg: CREATE BRANCH SQL implementation

[zhangbutao]

What changes were proposed in this pull request?

This PR refers to spark-sql about iceberg branch ddl implementation apache/iceberg#6617
If someone has different opinions about the sql syntax of branch, we can discuss here.

Why are the changes needed?

Personally, branch is more useful than snapshot in iceberg, and it is more friendly to users. We can use branch do lots of meaningfull work.

Does this PR introduce any user-facing change?

Added a new sql syntax and hive users can create iceberg branch using the sql.

ALTER TABLE tableName
{CREATE BRANCH branchName [FOR SYSTEM_VERSION AS OF {snapshotId} | FOR SYSTEM_TIME AS OF {timestamp}]
[RETAIN interval {DAYS | HOURS | MINUTES}]
[WITH SNAPSHOT RETENTION {[num_snapshots SNAPSHOTS] [interval {DAYS | HOURS | MINUTES}]}]}]

How was this patch tested?

UT

[aturoczy]

uhhhhh! This is super cool idea! I want it! Actually with this it is possible to play on dataset without any risk. I love it! I want it!
@deniskuzZ@ayushtkn@simhadri-g Could you please make this review a priority?

Documents about it:
https://docs.google.com/document/d/1tbATFPrKF3vNlzkgZQdaW8CAJmbjvryfrlg6C2Ci_aA/edit#heading=h.v8gsu2fe19q2

This is my new favorite PR (After the Docker support)

[ayushtkn]

Didn't check the code, But it was something I was also reading coincidently 2-3 hours before only :)

The main thing to chase is not creating a branch. But to insert into those branches. @zhangbutao We should chase that once we have the create in. Will review this in a day or two along with Denys. :)

Just FYI. I think Iceberg has options to create Tags as well now, and an option to do cherry-pick to branches as well. They are gonna release it for spark in 1.12.0

Good to have stuff!!!

Was reading here:
https://www.dremio.com/blog/exploring-branch-tags-in-apache-iceberg-using-spark/

[aturoczy]

Tags!!!! 😍

[aturoczy]

[zhangbutao]

@TuroczyX @ayushtkn yes, branch and tag are important feats in iceberg 1.2.0. I have created a umbrella ticket https://issues.apache.org/jira/browse/HIVE-27233 to track the new feats about branch&tag.

The main thing to chase is not creating a branch. But to insert into those branches.

I absolutely agree. I am exploring how to achieve this in hive. Next maybe we shoud consider upgrade iceberg to latest version 1.2.1. In addition, we can refer to this PR apache/iceberg#6965 to implement the sql syntax for insert into branch

[zhangbutao]

#4252 Upgrade iceberg to 1.2.1 in order to better integrate branch&tag features.

[aturoczy]

yes, let's upgrade the iceberg version

[deniskuzZ]

Thanks for the path @zhangbutao! LGTM, some minor. comments, please take a look once you have a chance. Thank you!

[deniskuzZ]

LGTM +1 (1 minor comment), pending tests

[zhangbutao]

Git rebase to fix code conflicts.

[sonarqubecloud]

Kudos, SonarCloud Quality Gate passed!

2 Bugs
0 Vulnerabilities
0 Security Hotspots
11 Code Smells

No Coverage information
No Duplication information

[zhangbutao]

@deniskuzZ@ayushtkn Can we merge this PR first? Thanks.