refine range representation on the result of explain

[zz-jason]

Feature Request

At present, the range representation is a little confusing:

TiDB(root@127.0.0.1:test) > explain select * from t use index(idx);
+-------------------+-------+------+---------------------------------------------------------------------+
| id                | count | task | operator info                                                       |
+-------------------+-------+------+---------------------------------------------------------------------+
| IndexReader_5     | 2.00  | root | index:IndexScan_4                                                   |
| └─IndexScan_4     | 2.00  | cop  | table:t, index:a, range:[NULL,+inf], keep order:false, stats:pseudo |
+-------------------+-------+------+---------------------------------------------------------------------+
2 rows in set (0.00 sec)

TiDB(root@127.0.0.1:test) > explain select * from t use index(idx) where a is not null;
+-------------------+-------+------+---------------------------------------------------------------------+
| id                | count | task | operator info                                                       |
+-------------------+-------+------+---------------------------------------------------------------------+
| IndexReader_6     | 2.00  | root | index:IndexScan_5                                                   |
| └─IndexScan_5     | 2.00  | cop  | table:t, index:a, range:[-inf,+inf], keep order:false, stats:pseudo |
+-------------------+-------+------+---------------------------------------------------------------------+
2 rows in set (0.00 sec)

as you can see in the above:

• a full range are expressed with [NULL,+inf]
• a range without any NULL value is expressed with [-inf,+inf]

Describe the feature you'd like:

Can we regard NULL as the smallest value:

• [NULL,+inf) to express a full range
• (NULL, +inf)to express a range without NULL value

Describe alternatives you've considered:

No

Teachability, Documentation, Adoption, Migration Strategy:

No

[zz-jason]

It's the same issue with #9231. For the same reason #9231 (comment), I'm going to close it as well.