feat(storage): Support domain for nested data types

[b41sh]

I hereby agree to the terms of the CLA available at: https://databend.rs/dev/policies/cla/

Summary

• Store max/min values for all the leaf columns of nested data types
• Generate domain for nested data types and skip the block if the filter range does not match the domain

From local test data, it can effectively reduce the query time when the filter ranges do not match

MySQL root@127.0.0.1:default> create table tt (a array(int), b map(string, int), c tuple(int, string));
Query OK, 0 rows affected
Time: 0.223s
MySQL root@127.0.0.1:default> select count(*) from tt;
+----------+
| count(*) |
+----------+
| 100000   |
+----------+
1 row in set
Time: 0.087s
MySQL root@127.0.0.1:default> select * from tt where a[1] >= 100;
+-------------------+---------------+----------------+
| a                 | b             | c              |
+-------------------+---------------+----------------+
| [100,101,102,103] | {'ip':99999}  | (100009,'abc') |
| [100,101,102,103] | {'ip':100000} | (100010,'abc') |
+-------------------+---------------+----------------+
2 rows in set
Time: 0.705s
MySQL root@127.0.0.1:default> select * from tt where a[1] >= 200;
+---+---+---+
| a | b | c |
+---+---+---+
+---+---+---+
0 rows in set
Time: 0.118s
MySQL root@127.0.0.1:default> select * from tt where b['ip'] >= 99999;
+-------------------+---------------+----------------+
| a                 | b             | c              |
+-------------------+---------------+----------------+
| [100,101,102,103] | {'ip':99999}  | (100009,'abc') |
| [100,101,102,103] | {'ip':100000} | (100010,'abc') |
+-------------------+---------------+----------------+
2 rows in set
Time: 0.792s
MySQL root@127.0.0.1:default> select * from tt where b['ip'] > 100000;
+---+---+---+
| a | b | c |
+---+---+---+
+---+---+---+
0 rows in set
Time: 0.108s
MySQL root@127.0.0.1:default> select * from tt where c.1 >= 100009;
+-------------------+---------------+----------------+
| a                 | b             | c              |
+-------------------+---------------+----------------+
| [100,101,102,103] | {'ip':99999}  | (100009,'abc') |
| [100,101,102,103] | {'ip':100000} | (100010,'abc') |
+-------------------+---------------+----------------+
2 rows in set
Time: 0.763s
MySQL root@127.0.0.1:default> select * from tt where c.1 >= 200009;
+---+---+---+
| a | b | c |
+---+---+---+
+---+---+---+
0 rows in set
Time: 0.116s

Closes #10487

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

1 Ignored Deployment

Name	Status	Preview	Comments	Updated
databend	⬜️ Ignored (Inspect)	Visit Preview		Mar 17, 2023 at 11:15AM (UTC)