我有一张桌子，我有一个人员列表，假设我有100个人列在该表中
我需要过滤掉使用不同标准的人并将它们分组，问题是当我开始排除第4至第5级时，性能问题出现并且变得缓慢
with lst_tous_movements as (
select
t1.refid_eClinibase
t1.[dthrfinmouvement]
t1.[unite_service_id]
t1.[unite_service_suiv_id]
from sometable t1
)
,lst_patients_hospitalisés as (
select distinct
t1.refid_eClinibase
from lst_tous_movements t1
where
t1.[dthrfinmouvement] = '4000-01-01'
)
,lst_patients_admisUIB_transferes as (
select distinct
t1.refid_eClinibase
from lst_tous_movements t1
left join lst_patients_hospitalisés t2 on t1.refid_eClinibase = t2.refid_eClinibase
where
t1.[unite_service_id] = 4
and t1.[unite_service_suiv_id] <> 0
and t2.refid_eClinibase is null
)
,lst_patients_admisUIB_nonTransferes as (
select distinct
t1.refid_eClinibase
from lst_tous_movements t1
left join lst_patients_admisUIB_transferes t2 on t1.refid_eClinibase = t2.refid_eClinibase
left join lst_patients_hospitalisés t3 on t1.refid_eClinibase = t3.refid_eClinibase
where
t1.[unite_service_id] = 4
and t1.[unite_service_suiv_id] = 0
and t2.refid_eClinibase is null
and t3.refid_eClinibase is null
)
,lst_patients_autres as (
select distinct
t1.refid_eClinibase
from lst_patients t1
left join lst_patients_admisUIB_transferes t2 on t1.refid_eClinibase = t2.refid_eClinibase
left join lst_patients_hospitalisés t3 on t1.refid_eClinibase = t3.refid_eClinibase
left join lst_patients_admisUIB_nonTransferes t4 on t1.refid_eClinibase = t4.refid_eClinibase
where
t2.refid_eClinibase is null
and t3.refid_eClinibase is null
and t4.refid_eClinibase is null
)
你可以看到我在这里进行了多级过滤......
当我到达第4级时，我的查询需要6到10秒才能完成
有什么方法可以加快速度吗？
这是我正在使用的数据集：
+------------------+-------------------------------+------------------+------------------+-----------------------+
| refid_eClinibase | nodossierpermanent_eClinibase | dthrfinmouvement | unite_service_id | unite_service_suiv_id |
+------------------+-------------------------------+------------------+------------------+-----------------------+
| 25611 | P0017379 | 2013-04-27 | 58 | 0 |
| 25611 | P0017379 | 2013-05-02 | 4 | 2 |
| 25611 | P0017379 | 2013-05-18 | 2 | 0 |
| 85886 | P0077918 | 2013-04-10 | 58 | 0 |
| 85886 | P0077918 | 2013-05-06 | 6 | 12 |
| 85886 | P0077918 | 4000-01-01 | 12 | 0 |
| 91312 | P0083352 | 2013-07-24 | 3 | 14 |
| 91312 | P0083352 | 2013-07-24 | 14 | 3 |
| 91312 | P0083352 | 2013-07-30 | 3 | 8 |
| 91312 | P0083352 | 4000-01-01 | 8 | 0 |
| 93835 | P0085879 | 2013-04-30 | 58 | 0 |
| 93835 | P0085879 | 2013-05-07 | 4 | 2 |
| 93835 | P0085879 | 2013-05-16 | 2 | 0 |
| 93835 | P0085879 | 2013-05-22 | 58 | 0 |
| 93835 | P0085879 | 2013-05-24 | 4 | 0 |
| 93835 | P0085879 | 2013-05-31 | 58 | 0 |
| 93836 | P0085880 | 2013-05-20 | 58 | 0 |
| 93836 | P0085880 | 2013-05-22 | 4 | 2 |
| 93836 | P0085880 | 2013-05-31 | 2 | 0 |
| 97509 | P0089576 | 2013-04-09 | 58 | 0 |
| 97509 | P0089576 | 2013-04-11 | 4 | 0 |
| 102787 | P0094886 | 2013-04-08 | 58 | 0 |
| 102787 | P0094886 | 2013-04-11 | 4 | 2 |
| 102787 | P0094886 | 2013-05-21 | 2 | 0 |
| 103029 | P0095128 | 2013-04-04 | 58 | 0 |
| 103029 | P0095128 | 2013-04-10 | 4 | 1 |
| 103029 | P0095128 | 2013-05-03 | 1 | 0 |
| 103813 | P0095922 | 2013-07-02 | 58 | 0 |
| 103813 | P0095922 | 2013-07-03 | 4 | 6 |
| 103813 | P0095922 | 2013-08-14 | 6 | 0 |
| 105106 | P0097215 | 2013-08-09 | 58 | 0 |
| 105106 | P0097215 | 2013-08-13 | 4 | 0 |
| 105106 | P0097215 | 2013-08-14 | 58 | 0 |
| 105106 | P0097215 | 4000-01-01 | 4 | 0 |
| 106223 | P0098332 | 2013-06-11 | 1 | 0 |
| 106223 | P0098332 | 2013-08-01 | 58 | 0 |
| 106223 | P0098332 | 4000-01-01 | 1 | 0 |
| 106245 | P0098354 | 2013-04-02 | 58 | 0 |
| 106245 | P0098354 | 2013-05-24 | 58 | 0 |
| 106245 | P0098354 | 2013-05-29 | 4 | 1 |
| 106245 | P0098354 | 2013-07-12 | 1 | 0 |
| 106280 | P0098389 | 2013-04-07 | 58 | 0 |
| 106280 | P0098389 | 2013-04-09 | 4 | 0 |
| 106416 | P0098525 | 2013-04-19 | 58 | 0 |
| 106416 | P0098525 | 2013-04-23 | 4 | 0 |
| 106444 | P0098553 | 2013-04-22 | 58 | 0 |
| 106444 | P0098553 | 2013-04-25 | 4 | 0 |
| 106609 | P0098718 | 2013-05-08 | 58 | 0 |
| 106609 | P0098718 | 2013-05-10 | 4 | 11 |
| 106609 | P0098718 | 2013-07-24 | 11 | 12 |
| 106609 | P0098718 | 4000-01-01 | 12 | 0 |
| 106616 | P0098725 | 2013-05-09 | 58 | 0 |
| 106616 | P0098725 | 2013-05-09 | 4 | 1 |
| 106616 | P0098725 | 2013-07-27 | 1 | 0 |
| 106698 | P0098807 | 2013-05-16 | 58 | 0 |
| 106698 | P0098807 | 2013-05-22 | 4 | 6 |
| 106698 | P0098807 | 2013-06-14 | 6 | 1 |
| 106698 | P0098807 | 2013-06-28 | 1 | 0 |
| 106714 | P0098823 | 2013-05-20 | 58 | 0 |
| 106714 | P0098823 | 2013-05-21 | 58 | 0 |
| 106714 | P0098823 | 2013-05-24 | 58 | 0 |
| 106729 | P0098838 | 2013-05-21 | 58 | 0 |
| 106729 | P0098838 | 2013-05-23 | 4 | 1 |
| 106729 | P0098838 | 2013-06-03 | 1 | 0 |
| 107038 | P0099147 | 2013-06-25 | 58 | 0 |
| 107038 | P0099147 | 2013-06-28 | 4 | 1 |
| 107038 | P0099147 | 2013-07-04 | 1 | 0 |
| 107038 | P0099147 | 2013-08-13 | 58 | 0 |
| 107038 | P0099147 | 2013-08-15 | 4 | 6 |
| 107038 | P0099147 | 4000-01-01 | 6 | 0 |
| 107082 | P0099191 | 2013-06-29 | 58 | 0 |
| 107082 | P0099191 | 2013-07-04 | 4 | 6 |
| 107082 | P0099191 | 2013-07-19 | 6 | 0 |
| 107157 | P0099267 | 4000-01-01 | 13 | 0 |
| 107336 | P0099446 | 4000-01-01 | 6 | 0 |
+------------------+-------------------------------+------------------+------------------+-----------------------+
感谢。
答案 0 :(得分：0)
很难准确理解问题的所有规则是什么，但一般的方法应该是在单个查询中添加“分组”列，该查询使用CASE语句对人员进行分类。 / p>
CASE中的条件按顺序进行评估，因此如果满足第一个条件，则甚至不会对该行评估后续条件。
以下是一些可以帮助您入门的代码....
select t1.refid_eClinibase
,t1.[dthrfinmouvement]
,t1.[unite_service_id]
,t1.[unite_service_suiv_id]
CASE WHEN [dthrfinmouvement] = '4000-01-01' THEN 'Group1 Label'
WHEN condition2 = something THEN 'Group2 Label'
....
WHEN conditionN = something THEN 'GroupN Label'
ELSE 'Catch All Label'
END as person_category
from sometable t1