【发布时间】：2018-05-11 15:12:47
【问题描述】：
我有以下 CTE 来拉入会员资格，然后拉入这些会员的所有索赔（如果他们在日期参数内有索赔），现在从总会员资格 (=1961) 我需要拉入 TOP索赔 CTE 的 3% 或 0.03。我看到 1961*0.03 舍入为 59，因此我需要从 Claims 中提取最高 (59) 个 Medicaid ID，其中索赔使用率最高。
例如，在 number_to_pull CTE 中给出需要拉入的线路（它给出了 3% 的成员资格），然后在 sum_of_claims CTE 中我想要仅从 Claims CTE 中提取前 3% 的 Medicaid ID。由于成员资格可能会根据日期参数而变化，因此我希望 sum_of_claims 具有以下内容，但我不知道如何开始
最终结果是，我将列出在调用日期范围内每次索赔点击次数最多的 Medicaid ID 列表前列 (3%)
我需要这样的东西，但我希望它提取 number_to_pull CTE 中的任何数字，然后根据索赔总和提取该数字。
Select Top ( select
round(count(mt.medicaid_no)*0.03) as percentt
from membership mt)
cll.medicaid_no
,count(distinct claim_number) as sum_of_claims
from claims cll
Group by cll.medicaid_no
) select * from sum_of_claims
这就是我的代码实际的样子
WITH
DATES AS
(
select TRUNC(TRUNC(SYSDATE,'y')-1,'y') as startdate,
TRUNC(SYSDATE,'y')-1 as enddate
from dual
),
membership as (
select Distinct
mbr.medicaid_no
,mbd.memb_dim_id
,mbd.memb_demographics_full_date
from dw.fact_member_demographics mbd
inner join dates d
on 1=1
inner join dw.DIM_MEMBER mbr
on mbd.memb_dim_id = mbr.memb_dim_id
Where EXTRACT(YEAR FROM mbd.memb_demographics_full_date)= extract(year from d.startdate)
and mbd.company_dim_id in ('575')
and mbd.age > 18
) ---select * from membership
,number_to_pull as (
select
round(count(mt.medicaid_no)*0.03) as percentt
from membership mt
) ---select * from top_number
,Claims as (
select
mbdd.medicaid_no
,mbdd.memb_dim_id
,dc.company_desc
,cl.primary_svc_date
,cl.claim_number
,case when cl.io_flag_dim_id = '1' then 'Inpatient'
when cl.io_flag_dim_id = '2' then 'Outpatient' else 'false' end as In_Op
,cl.admit_type
,proc.procedure_code
,dx1.diagnosis_code as dx1
,dx1.diagnosis_short_desc as dx1desc
,dx2.diagnosis_code as dx2
,dx2.diagnosis_short_desc as dx2desc
,dx3.diagnosis_code as dx3
,dx3.diagnosis_short_desc as dx3desc
,dx4.diagnosis_code as dx4
,dx4.diagnosis_short_desc as dx4desc
,dx5.diagnosis_code as dx5
,dx5.diagnosis_short_desc as dx5desc
,bt.inp_outp_ind
from membership mbdd
left join dw.fact_claim cl
on mbdd.memb_dim_id = cl.memb_dim_id
inner join dates d
on 1=1
inner join dw.DIM_PROCEDURE_CODE proc
on cl.cpt_code_dim_id = proc.procedure_dim_id
inner join dw.DIM_DIAGNOSIS dx1
on cl.diagnosis_1_dim_id = dx1.diagnosis_dim_id
inner join dw.DIM_DIAGNOSIS dx2
on cl.diagnosis_2_dim_id = dx2.diagnosis_dim_id
inner join dw.DIM_DIAGNOSIS dx3
on cl.diagnosis_3_dim_id = dx3.diagnosis_dim_id
inner join dw.DIM_DIAGNOSIS dx4
on cl.diagnosis_4_dim_id = dx4.diagnosis_dim_id
inner join dw.DIM_DIAGNOSIS dx5
on cl.diagnosis_5_dim_id = dx5.diagnosis_dim_id
inner join dw.DIM_BILL_TYPE bt
on cl.bill_type_dim_id = bt.bill_type_dim_id
inner join dw.DIM_COMPANY dc
on cl.company_dim_id = dc.company_dim_id
Where cl.primary_svc_date between d.startdate and d.enddate
and cl.company_dim_id in ('575')
and CL.WHOLE_CLAIM_STATUS_DIM_ID IN (1,2)
and cl.io_flag_dim_id in ('1','2')
) ---select * from claims
,sum_of_claims AS (
Select
---- this is where I want to pull in the top 3% based off of membeship and sum of claims per Medicaid
cll.medicaid_no
,count(distinct claim_number) as sum_of_claims
from claims cll
Group by cll.medicaid_no
) select * from sum_of_claims
我想要的最终结果是一份 Medicaid ID 列表和索赔总和，但这个列表只会是前 59 行 (3%)
MEDICAID_NO SUM_OF_CLAIMS
111111 $12,439.61
333333 $5,315.57
444444 $2,007.00
555555 $1,823.98
888888 $1,770.00
777777 $1,211.47
9999999 $1,157.61
6666666 $1,068.76
【问题讨论】：
-
样本数据和期望的结果真的很有帮助。
-
添加了基于语法的Oracle标签
-
谢谢。我刚刚更新了最终结果。
-
CTE 代表什么？
-
@oreh Common Table Expression，即
with子句。
标签： sql oracle common-table-expression percentage