no_ Expand and use_ concat

RDBMS： Database Release Update Revision : 19.11.2.0.211019 (33153989)

Reference documents ：

(4 Bar message ) Query conversion or expansion_ Document Porter -CSDN Blog

Query Transformations (oracle.com)

In OR expansion, the optimizer transforms a query block containing top-level disjunctions into the form of a UNION ALL query that contains two or more branches. The optimizer achieves this goal by splitting the disjunction into its components, and then associating each component with a branch of a UNION ALL query.

The optimizer can choose OR expansion for various reasons. For example, it may enable more efficient access paths or alternative join methods that avoid Cartesian products. As always, the optimizer performs the expansion only if the cost of the transformed statement is lower than the cost of the original statement.

In previous releases, the optimizer used the CONCATENATION operator to perform the OR expansion. Starting in Oracle Database 12c Release 2 (12.2), the optimizer uses the UNION-ALL operator instead. The framework provides the following enhancements:

• Enables interaction among various transformations

• Avoids sharing query structures

• Enables the exploration of various search strategies

• Provides the reuse of cost annotation

• Supports the standard SQL syntax

-- About these two hint Explanation  

USE_CONCAT -  The USE_CONCAT hint forces combined OR conditions in the WHERE clause of a query to be transformed into a compound query using the UNION ALL set operator. Generally, this transformation occurs only if the cost of the query using the concatenations is cheaper than the cost without them.

NO_EXPAND - The NO_EXPAND hint prevents the cost-based optimizer from considering OR-expansion for queries having OR conditions or IN-lists in the WHERE clause. Usually, the optimizer considers using OR expansion and uses this method if it decides that the cost is lower than not using it.

-- test ：

-- where There is... In the condition or, The optimizer uses union all 

[email protected]>SELECT *
  2  FROM   hr.employees e, hr.departments d
  3  WHERE  (e.email='SSTILES' OR d.department_name='Treasury')
  4  AND    e.department_id = d.department_id;


Execution Plan
----------------------------------------------------------
Plan hash value: 2462079978

----------------------------------------------------------------------------------------------------
| Id  | Operation                      | Name              | Rows  | Bytes | Cost (%CPU)| Time     |
----------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT               |                   |    11 |  2079 |     6   (0)| 00:00:01 |
|   1 |  VIEW                          | VW_ORE_19FF4E3E   |    11 |  2079 |     6   (0)| 00:00:01 |
|   2 |   UNION-ALL                    |                   |       |       |            |          |
|   3 |    NESTED LOOPS                |                   |     1 |    90 |     2   (0)| 00:00:01 |
|   4 |     TABLE ACCESS BY INDEX ROWID| EMPLOYEES         |     1 |    69 |     1   (0)| 00:00:01 |
|*  5 |      INDEX UNIQUE SCAN         | EMP_EMAIL_UK      |     1 |       |     0   (0)| 00:00:01 |
|   6 |     TABLE ACCESS BY INDEX ROWID| DEPARTMENTS       |     1 |    21 |     1   (0)| 00:00:01 |
|*  7 |      INDEX UNIQUE SCAN         | DEPT_ID_PK        |     1 |       |     0   (0)| 00:00:01 |
|   8 |    NESTED LOOPS                |                   |    10 |   900 |     4   (0)| 00:00:01 |
|   9 |     NESTED LOOPS               |                   |    10 |   900 |     4   (0)| 00:00:01 |
|* 10 |      TABLE ACCESS FULL         | DEPARTMENTS       |     1 |    21 |     3   (0)| 00:00:01 |
|* 11 |      INDEX RANGE SCAN          | EMP_DEPARTMENT_IX |    10 |       |     0   (0)| 00:00:01 |
|* 12 |     TABLE ACCESS BY INDEX ROWID| EMPLOYEES         |    10 |   690 |     1   (0)| 00:00:01 |
----------------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   5 - access("E"."EMAIL"='SSTILES')
   7 - access("E"."DEPARTMENT_ID"="D"."DEPARTMENT_ID")
  10 - filter("D"."DEPARTMENT_NAME"='Treasury')
  11 - access("E"."DEPARTMENT_ID"="D"."DEPARTMENT_ID")
  12 - filter(LNNVL("E"."EMAIL"='SSTILES'))

Note
-----
   - this is an adaptive plan


Statistics
----------------------------------------------------------
        156  recursive calls
          0  db block gets
        146  consistent gets
          9  physical reads
          0  redo size
       1760  bytes sent via SQL*Net to client
        510  bytes received via SQL*Net from client
          2  SQL*Net roundtrips to/from client
          7  sorts (memory)
          0  sorts (disk)
          1  rows processed

[email protected]>

 -- add hint  no_expand , Found that he didn't go union all 了 . from Statistics Look up , Cycle call , Consistent read , Sort , A lot less .

SELECT /*+ NO_EXPAND */ *
FROM   hr.employees e, hr.departments d
WHERE  (e.email='SSTILES' OR d.department_name='Treasury')
AND    e.department_id = d.department_id;


[email protected]>SELECT /*+ NO_EXPAND */ *
  2  FROM   hr.employees e, hr.departments d
  3  WHERE  (e.email='SSTILES' OR d.department_name='Treasury')
  4  AND    e.department_id = d.department_id;


Execution Plan
----------------------------------------------------------
Plan hash value: 413710308

--------------------------------------------------------------------------------------------
| Id  | Operation                    | Name        | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT             |             |     5 |   450 |     6  (17)| 00:00:01 |
|   1 |  MERGE JOIN                  |             |     5 |   450 |     6  (17)| 00:00:01 |
|   2 |   TABLE ACCESS BY INDEX ROWID| DEPARTMENTS |    27 |   567 |     2   (0)| 00:00:01 |
|   3 |    INDEX FULL SCAN           | DEPT_ID_PK  |    27 |       |     1   (0)| 00:00:01 |
|*  4 |   FILTER                     |             |       |       |            |          |
|*  5 |    SORT JOIN                 |             |   107 |  7383 |     4  (25)| 00:00:01 |
|   6 |     TABLE ACCESS FULL        | EMPLOYEES   |   107 |  7383 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   4 - filter("E"."EMAIL"='SSTILES' OR "D"."DEPARTMENT_NAME"='Treasury')
   5 - access("E"."DEPARTMENT_ID"="D"."DEPARTMENT_ID")
       filter("E"."DEPARTMENT_ID"="D"."DEPARTMENT_ID")


Statistics
----------------------------------------------------------
          8  recursive calls
          0  db block gets
         12  consistent gets
          0  physical reads
          0  redo size
       1760  bytes sent via SQL*Net to client
        527  bytes received via SQL*Net from client
          2  SQL*Net roundtrips to/from client
          1  sorts (memory)
          0  sorts (disk)
          1  rows processed

[email protected]>

END