diff: can use multiple columns to split chunks

[WangXiangUSTC]

What problem does this PR solve?

1. use tidb's statistical information to split chunk
2. if no tidb's statistical information, use random function to split chunks, and can use multiple columns to split.

What is changed and how it works?

1. for use tidb's statistical information, we can get buckets info by execute "show STATS_BUCKETS"
   for example:

mysql> SHOW STATS_BUCKETS WHERE db_name= "test" AND table_name="testa";
 			+---------+------------+----------------+-------------+----------+-----------+-------+---------+---------------------+---------------------+
 			| Db_name | Table_name | Partition_name | Column_name | Is_index | Bucket_id | Count | Repeats | Lower_Bound         | Upper_Bound         |
 			+---------+------------+----------------+-------------+----------+-----------+-------+---------+---------------------+---------------------+
 			| test    | testa      |                | PRIMARY     |        1 |         0 |    64 |       1 | 1846693550524203008 | 1846838686059069440 |
 			| test    | testa      |                | PRIMARY     |        1 |         1 |   128 |       1 | 1846840885082324992 | 1847056389361369088 |
 			+---------+------------+----------------+-------------+----------+-----------+-------+---------+---------------------+---------------------+

we can get the lower_bound and upper_bound as chunk's range.

2. for random function, for example, there are these datas in mysql:

+------+------+
| id   | name |
+------+------+
|    1 | a    |
|    1 | b    |
|    1 | c    |
|    2 | c    |
|    2 | d    |
|    2 | c    |
+------+------+

will split to these chunks when chunk size is 1:

`id` = 1 AND `name` = "a"
`id` = 1 AND `name` > "a" AND `name` < "b"
`id` = 1 AND `name` >= "b" AND `name` < "c" 
`id` = 1 AND `name` = "c" 
`id` = 1 AND `name` > "c" 
`id` = 1 AND `name` < "a"
`id` > 1 AND `id` < 2 AND 
`id` = 2 AND `name` = "c" 
`id` = 2 AND `name` > "c" AND `name` <= "d" 
`id` = 2 AND `name` > "d" 
`id` = 2 AND `name` < "c"
`id` > 2 
`id` < 1 

Check List

Tests

• Manual test

Related changes

• Need to update the documentation
• Need to be included in the release note

[IANTHEREAL]

CI was broken

[IANTHEREAL]

good job！

[IANTHEREAL]

@kennytm PTAL

[IANTHEREAL]

LGTM

[csuzhangxc]

is it possible one column appended multi times?

[WangXiangUSTC]

no, a index can only contain unique column

[WangXiangUSTC]

@csuzhangxc PTAL again

[csuzhangxc]

LGTM. but you need to resolve the conflict now.

[WangXiangUSTC]

thanks @csuzhangxc