我在表上应用row_number（）（通过varchar2列的顺序）有200万条记录，但在oracle 12c中花费的时间更多。
计划哈希值：193104477
-----------------------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes |TempSpc| Cost (%CPU)| Time |
-----------------------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1631K| 628M| | 257K (1)| 00:00:11 |
| 1 | WINDOW SORT | | 1631K| 628M| 670M| 257K (1)| 00:00:11 |
|* 2 | HASH JOIN | | 1631K| 628M| 8128K| 117K (1)| 00:00:05 |
| 3 | TABLE ACCESS FULL | SYS_USER | 108K| 6856K| | 2466 (1)| 00:00:01 |
|* 4 | HASH JOIN | | 1631K| 527M| 2608K| 87280 (1)| 00:00:04 |
| 5 | TABLE ACCESS FULL | CHANGE_REQUEST | 45210 | 2075K| | 3850 (1)| 00:00:01 |
|* 6 | HASH JOIN | | 1632K| 454M| 3800K| 59840 (1)| 00:00:03 |
| 7 | TABLE ACCESS FULL | SYS_USER | 108K| 2531K| | 2467 (1)| 00:00:01 |
|* 8 | HASH JOIN | | 1632K| 417M| 7800K| 35566 (1)| 00:00:02 |
|* 9 | HASH JOIN | | 62854 | 7058K| | 1826 (1)| 00:00:01 |
| 10 | INDEX FAST FULL SCAN | COMPINDEX_SYSID_NAME_NI | 2716 | 143K| | 15 (0)| 00:00:01 |
| 11 | VIEW | index$_join$_003 | 69097 | 4116K| | 1810 (1)| 00:00:01 |
|* 12 | HASH JOIN | | | | | | |
|* 13 | INDEX FAST FULL SCAN| I001823ASSIGNMENT_GROUPSYSAPPR | 69097 | 4116K| | 1008 (0)| 00:00:01 |
| 14 | INDEX FAST FULL SCAN| SYSAPPROVAL_GROUP_SYS_ID | 69097 | 4116K| | 743 (0)| 00:00:01 |
| 15 | TABLE ACCESS FULL | SYSAPPROVAL_APPROVER | 1841K| 268M| | 18984 (1)| 00:00:01 |
以下是我正在使用的查询。
SELECT cr.number_x AS UNIQUE_KEY,
row_number() over(ORDER BY sa.sys_id) AS COUNTER,
sa.sys_updated_on AS "DATE"
FROM snowdb.sysapproval_approver sa,
snowdb.change_request cr,
snowdb.sysapproval_group sag,
snowdb.sys_user_group sug,
snowdb.sys_user su,
snowdb.sys_user su1
WHERE sa."GROUP_X"=sag.sys_id
AND sa.sysapproval=cr.sys_id
AND sag.assignment_group=sug.sys_id
AND sa.approver=su.sys_id
AND sa.sys_updated_by=su1.employee_number