[#3282] improvement: Support sort order when create the Doris table.

[yuqi1129]

What changes were proposed in this pull request?

1. Move check sortorder logic into the corresponding implementation from JdbcTableOperation
2. Support SortOrder when create a Doris table.

Why are the changes needed?

Doris supports creating a sorted table, we need to implement it in Gravitino API.

Close: #3282

Does this PR introduce any user-facing change?

N/A.

How was this patch tested?

Add test testSortOrderTable

[yuqi1129]

@zhoukangcn Could you kindly review it for me?

[yuqi1129]

After discussion with @zhoukangcn , this PR should be holds until we confirm that the syntax of DUPLICATE KEY equals to Sort order. It seems that SR does not have this issue.

[yuqi1129]

After discussion with @zhoukangcn , this PR should be holds until we confirm that the syntax of DUPLICATE KEY equals to Sort order. It seems that SR does not have this issue.

@zhoukangcn According to the document from 2.0, the sort order syntax in Doirs is:

https://doris.apache.org/docs/2.0/table-design/index/prefix-index/

I'm wondering if we should proceed with this PR.

[zhoukangcn]

@yuqi1129 I had notice this Doris document when we discuss last time.

I believe this document only describes that data will be sorted and stored when using the Aggregate, Unique, and Duplicate data models.

In this document https://doris.apache.org/docs/1.2/sql-manual/sql-reference/Data-Definition-Statements/Create/CREATE-TABLE

- keys_type

Data model.

key_type(col1, col2, ...)

key_type supports the following models:

DUPLICATE KEY (default): The subsequent specified column is the sorting column.

AGGREGATE KEY: The specified column is the dimension column.

UNIQUE KEY: The subsequent specified column is the primary key column.

So, I think DUPLICATE, AGGREGATE, UNIQUE should be regarded as data models, not sort keys.

[yuqi1129]

doris.apache.org/docs/2.0/table-design/index/prefix-index

@yuqi1129 I had notice this Doris document when we discuss last time.

I believe this document only describes that data will be sorted and stored when using the Aggregate, Unique, and Duplicate data models.

In this document doris.apache.org/docs/1.2/sql-manual/sql-reference/Data-Definition-Statements/Create/CREATE-TABLE

- keys_type

Data model.

key_type(col1, col2, ...)

key_type supports the following models:

DUPLICATE KEY (default): The subsequent specified column is the sorting column.

AGGREGATE KEY: The specified column is the dimension column.

UNIQUE KEY: The subsequent specified column is the primary key column.

So, I think DUPLICATE, AGGREGATE, UNIQUE should be regarded as data models, not sort keys.

I see, there seems to be a big difference between version 1.2 and 2.X. In 1.2 DUPLICATE, AGGREGATE, UNIQUE are reviewed as data models, however in 2.X, as the documents described, it can also affect the data storage format and can be regarded as the sort key.

In all, we will postpone the PR until we have a full understanding of it.