Add StarRocks connector

[Max-Cheng]

Description

Support Trino access Starrocks by connector

Additional context and related issues

Release notes

( ) This is not user-visible or is docs only, and no release notes are required.
(x) Release notes are required. Please propose a release note for me.
( ) Release notes are required, with the following suggested text:

## Section
* Add Starrocks Connector

[ebyhr]

@hackeryang jdbc lable is for JDBC driver, not JDBC-based connectors.

[ebyhr]

The old PR #17330 looks better than this PR at a glance. I would recommend asking @chenjian2664 if the PR can be reopened and how to collaborate.

[Max-Cheng]

The old PR #17330 looks better than this PR at a glance. I would recommend asking @chenjian2664 if the PR can be reopened and how to collaborate.

@ebyhr In fact, these are two different implementations.

This PR using a JDBC client can first fetch the query plan from the responsible FE, then distribute the fetched query plan as a parameter to all relevant BEs, and finally fetch the data returned by the BEs.

PR #17330 using JDBC connector can only read data from a single FE at a time. Data reading is very slow.

You can take a look at https://docs.starrocks.io/docs/unloading/Flink_connector/

[Max-Cheng]

This is benchmark on Starrocks 3.2.9 Using JDBC connector and Starrocks connector.
Comparison test on TPC-H 100G scale dataset with 22 queries.

Test data

Table	Rows
customer	15 million
lineitem	600 million
nation	25
orders	150 million
part	20 million
partsupp	80 million
region	5
supplier	1 million

Result

Query	Trino-435 [SR connector->SR]	Trino-435 【mysql connector->SR】
Q1	24470	804226
Q2	13547	70248
Q3	17045	161474
Q4	18691	240217
Q5	22624	336967
Q6	661	3979
Q7	11909	153452
Q8	27098	440370
Q9	37751	505058
Q10	10074	87535
Q11	4340	83074
Q12	7140	115119
Q13	13516	129997
Q14	3613	14239
Q15	2433	13642
Q16	9974	45004
Q17	15635	473364
Q18	43138	210435
Q19	5076	21069
Q20	14661	70100
Q21	49115	355466
Q22	6306	11543
Total	359S	4347S

[ebyhr]

Please fix CI failures and add following tests:

• Test extending BaseConnectorTest
• Type mapping test. See TestPostgreSqlTypeMapping.
• A product test

Also, I would recommend reading developer guide.

[chenjian2664]

@Max-Cheng I love your implementation of the load data from BEs, great job! And I think the main difference is just between the SplitSource part(with jdbc implementation), actually it can envolve on the jdbc implementation.
For the performance, IIRC, my user case is pretty simple that user using point query with limit and the read performance is enough for the table is not so big doing data exploration.

@ebyhr Does the community still want the jdbc implementation of the Starrocks? if yes I would like to reopen that.