我正在尝试对一个非常大的数据集运行查询，虽然由于我在oracle中缺乏专业知识，我似乎无法加速它并且查询需要很长时间，因为有很多数据行。我附上解释计划，以显示究竟在做什么。任何帮助将不胜感激。
select
a.customer_id,
c.vm_mor,
c.vm_id,
d.license_id,
e.product,
f.guest_os_id
from
customers a,
vm_groups b,
vms c,
vm_licenses d,
licenses e,
vm_compute_histories f
where
a.customer_id = b.customer_id
and b.vm_group_id = c.vm_group_id
and c.vm_id = d.vm_id
and d.license_id = e.license_id
and c.vm_id = f.vm_id
order by
a.customer_id, vm_id
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Plan hash value: 3321561485
---------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
---------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 7575 | 591K| 36 (12)| 00:00:01 |
| 1 | SORT ORDER BY | | 7575 | 591K| 36 (12)| 00:00:01 |
|* 2 | HASH JOIN | | 7575 | 591K| 35 (9)| 00:00:01 |
|* 3 | HASH JOIN | | 1238 | 87898 | 17 (12)| 00:00:01 |
| 4 | TABLE ACCESS FULL | LICENSES | 45 | 1665 | 3 (0)| 00:00:01 |
|* 5 | HASH JOIN | | 1241 | 42194 | 14 (15)| 00:00:01 |
| 6 | VIEW | index$_join$_002 | 46 | 368 | 3 (34)| 00:00:01 |
|* 7 | HASH JOIN | | | | | |
| 8 | INDEX FAST FULL SCAN| XIF1VM_GROUPS | 46 | 368 | 1 (0)| 00:00:01 |
| 9 | INDEX FAST FULL SCAN| XPKVM_GROUPS | 46 | 368 | 1 (0)| 00:00:01 |
|* 10 | HASH JOIN | | 1241 | 32266 | 11 (10)| 00:00:01 |
| 11 | TABLE ACCESS FULL | VMS | 1062 | 18054 | 7 (0)| 00:00:01 |
| 12 | INDEX FAST FULL SCAN | XIE1VM_LICENSES | 1241 | 11169 | 3 (0)| 00:00:01 |
| 13 | TABLE ACCESS FULL | VM_COMPUTE_HISTORIES | 7780 | 70020 | 17 (0)| 00:00:01 |
---------------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("C"."VM_ID"="F"."VM_ID")
3 - access("D"."LICENSE_ID"="E"."LICENSE_ID")
5 - access("B"."VM_GROUP_ID"="C"."VM_GROUP_ID")
7 - access(ROWID=ROWID)
10 - access("C"."VM_ID"="D"."VM_ID")
29 rows selected.
以下是要求的解释计划：
这是查询：
SELECT a.customer_id,
c.vm_mor,
c.vm_id,
d.license_id,
e.product,
f.guest_os_id
FROM customers a,
vms c,
vm_licenses d,
licenses e,
vm_compute_histories f
WHERE EXISTS
(SELECT 'X'
FROM vm_groups b
WHERE a.customer_id = b.customer_id
AND c.vm_group_id = b.vm_group_id)
AND c.vm_id = d.vm_id
AND d.license_id = e.license_id
AND c.vm_id = f.vm_id
ORDER BY a.customer_id, c.vm_id
/
计划：
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Plan hash value: 1991801996
----------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 2354 | 144K| 31 (13)| 00:00:01 |
| 1 | SORT ORDER BY | | 2354 | 144K| 31 (13)| 00:00:01 |
|* 2 | HASH JOIN | | 2354 | 144K| 30 (10)| 00:00:01 |
| 3 | NESTED LOOPS | | 388 | 20952 | 15 (20)| 00:00:01 |
|* 4 | HASH JOIN | | 388 | 19400 | 15 (20)| 00:00:01 |
| 5 | SORT UNIQUE | | 46 | 368 | 3 (34)| 00:00:01 |
| 6 | VIEW | index$_join$_006 | 46 | 368 | 3 (34)| 00:00:01 |
|* 7 | HASH JOIN | | | | | |
| 8 | INDEX FAST FULL SCAN | XIF1VM_GROUPS | 46 | 368 | 1 (0)| 00:00:01 |
| 9 | INDEX FAST FULL SCAN | XPKVM_GROUPS | 46 | 368 | 1 (0)| 00:00:01 |
|* 10 | HASH JOIN | | 388 | 16296 | 12 (17)| 00:00:01 |
| 11 | MERGE JOIN | | 388 | 9700 | 5 (20)| 00:00:01 |
| 12 | TABLE ACCESS BY INDEX ROWID| LICENSES | 22 | 374 | 2 (0)| 00:00:01 |
| 13 | INDEX FULL SCAN | XPKLICENSES | 22 | | 1 (0)| 00:00:01 |
|* 14 | SORT JOIN | | 388 | 3104 | 3 (34)| 00:00:01 |
| 15 | INDEX FAST FULL SCAN | XIE1VM_LICENSES | 388 | 3104 | 2 (0)| 00:00:01 |
| 16 | TABLE ACCESS FULL | VMS | 953 | 16201 | 6 (0)| 00:00:01 |
|* 17 | INDEX UNIQUE SCAN | XPKCUSTOMERS | 1 | 4 | 0 (0)| 00:00:01 |
| 18 | TABLE ACCESS FULL | VM_COMPUTE_HISTORIES | 5769 | 51921 | 15 (0)| 00:00:01 |
----------------------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("C"."VM_ID"="F"."VM_ID")
4 - access("C"."VM_GROUP_ID"="B"."VM_GROUP_ID")
7 - access(ROWID=ROWID)
10 - access("C"."VM_ID"="D"."VM_ID")
14 - access("D"."LICENSE_ID"="E"."LICENSE_ID")
filter("D"."LICENSE_ID"="E"."LICENSE_ID")
17 - access("A"."CUSTOMER_ID"="B"."CUSTOMER_ID")
36 rows selected.
SQL>
答案 0 :(得分：2)
试试这个，它可能会加快你的查询速度
SELECT /*+ use_hash(a,b,c,d,e,f) */
在查询中使用此SELECT语句。
答案 1 :(得分：1)
尝试将解释计划与此SQL进行比较。我无法将其粘贴为评论，因此我必须将其作为＆＃34;回答＆＃34;发布。我从你的查询中注意到的是你从不投射表格中的值＆＃34; b＆＃34;所以EXISTS可能有助于使事情变得更快......
检查出来：
SELECT a.customer_id,
c.vm_mor,
c.vm_id,
d.license_id,
e.product,
f.guest_os_id
FROM customers a,
vms c,
vm_licenses d,
licenses e,
vm_compute_histories f
WHERE EXISTS
(SELECT 'X'
FROM vm_groups vm1
WHERE a.customer_id = vm.customer_id
AND c.vm_group_id = vm.vm_group_id)
AND c.vm_id = d.vm_id
AND d.license_id = e.license_id
AND c.vm_id = f.vm_id
ORDER BY a.customer_id, vm_id