[ISSUE-154] Support Array[Tuple] clickhouse type

[qingfei1994]

Why are the changes needed?

Close #154 .

Brief change log

• support Row type during serialization and deserialization

How was this patch tested?

• Add some test cases that check the changes thoroughly including negative and positive cases if possible

• Add screenshots for manual tests if appropriate

• Run test locally before making a pull request

Test case:
Clickhouse DDL

set flatten_nested=0;
create table test_tuple3(arr Array(Tuple(a Int32, b Int32, c Int32)), z Int32) Engine=MergeTree() Order By tuple();

Flink SQL

 CREATE TABLE test_tuple (
    arr Array<Row<a int,b int,c int>>,
    z int
  ) WITH (
    'connector' = 'clickhouse',
    'url' = '......',
    'database-name' = 'default',
    'table-name' = 'test_tuple3',
  .....
  );
  insert into test_tuple values(Array[(1,2,3)],700);

Query result

SELECT *
FROM test_tuple3
WHERE z = 700
   ┌─arr───────┬───z─┐
1. │ [(1,2,3)] │ 700 │
   └───────────┴─────┘

1 row in set. Elapsed: 0.004 sec.

Documentation

• Does this pull request introduce a new feature? (yes / no) no
• If yes, how is the feature documented? (not applicable / docs / JavaDocs / not documented)

[itinycheng]

Hi @qingfei1994

Good PR, but it will be better if it works well with ClickHouseCatalog.
Run the command ./mvnw spotless:apply before submitting can help us to ensure the code format is correct.

[qingfei1994]

Hi @qingfei1994

Good PR, but it will be better if it works well with ClickHouseCatalog. Run the command ./mvnw spotless:apply before submitting can help us to ensure the code format is correct.

Sure I've fix the checkstyle violation, and will also check if it works with ClickhouseCatalog

[qingfei1994]

Hi @qingfei1994
Good PR, but it will be better if it works well with ClickHouseCatalog. Run the command ./mvnw spotless:apply before submitting can help us to ensure the code format is correct.

Sure I've fix the checkstyle violation, and will also check if it works with ClickhouseCatalog

@itinycheng @czy006
I've tested and fixed the ClickhouseCatalog with Row data type.
Flink SQL:

CREATE catalog clickhouse_catalog WITH (
  'type' = 'clickhouse',
  'url' = '.....',
  'database-name' = 'test',
  'username' = '',
  'password' = ''
);

create table print_sink
(
    arr Array<Row<a int,b int, c int>>,
    z int
) with (
'connector'='print'
);

insert into print_sink
select
  arr,
  z
from
  clickhouse_catalog.`default`.test_tuple3;

query result:

2> +I[[+I[1, 2, 3]], 600]
2> +I[[+I[1, 2, 3]], 600]
2> +I[[+I[4, 5, 6]], 800]
2> +I[[+I[4, 5, 6]], 800]
2> +I[[+I[1, 2, 3]], 700]

please check the latest submitted code.

[itinycheng]

Hi @qingfei1994

I have merged this PR.
One more thing, tuple elements without names are not included in
the test cases above, like ‘Tuple(T1, T2, ...)’