我正在尝试计算度量的中位数，但我需要将上一年/月/工作日的中位数数据显示在当前数据上。
date wday admns median result
28/10/2016 Friday 7 37.5 0
21/10/2016 Friday 34 37.5 0
7/10/2016 Friday 41 37.5 0
14/10/2016 Friday 63 37.5 0
20/10/2017 Friday 23 49 37.5
27/10/2017 Friday 44 49 37.5
6/10/2017 Friday 54 49 37.5
13/10/2017 Friday 63 49 37.5
在结果栏中，我需要将上一年/月/日的中位数显示到当前年份。对于2015年，没有数据，因此2016年数据的结果字段为0。
我使用此查询来获取中位数
with a as (select dt,weekday,admns,extract(year from dt) year,to_char(trunc(sysdate,'MM'),'Month') month from med_drop)
select dt, weekday, admns , year,month,
median(admns) over ( partition by year,weekday) export_table
from a;
测试数据
CREATE TABLE export_table
(DT DATE,
WEEKDAY VARCHAR2(10),
ADMNS NUMBER
);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('01/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Sunday',10);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('02/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Monday',12);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('03/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Tuesday',15);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('04/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Wednesday',52);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('05/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Thursday',23);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('06/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Friday',54);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('07/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Saturday',78);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('08/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Sunday',66);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('09/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Monday',5);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('10/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Tuesday',4);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('11/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Wednesday',78);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('12/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Thursday',99);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('13/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Friday',63);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('14/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Saturday',71);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('15/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Sunday',63);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('16/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Monday',89);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('17/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Tuesday',23);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('18/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Wednesday',45);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('19/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Thursday',87);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('20/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Friday',23);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('21/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Saturday',12);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('22/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Sunday',11);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('23/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Monday',63);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('24/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Tuesday',93);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('25/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Wednesday',23);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('26/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Thursday',88);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('27/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Friday',44);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('28/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Saturday',441);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('29/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Sunday',66);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('30/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Monday',23);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('31/OCT/2017 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Tuesday',16);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('01/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Saturday',55);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('02/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Sunday',67);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('03/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Monday',41);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('04/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Tuesday',85);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('05/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Wednesday',25);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('06/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Thursday',63);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('07/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Friday',41);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('08/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Saturday',87);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('09/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Sunday',46);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('10/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Monday',14);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('11/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Tuesday',74);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('12/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Wednesday',36);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('13/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Thursday',74);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('14/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Friday',63);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('15/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Saturday',14);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('16/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Sunday',96);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('17/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Monday',31);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('18/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Tuesday',41);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('19/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Wednesday',23);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('20/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Thursday',96);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('21/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Friday',34);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('22/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Saturday',74);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('23/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Sunday',63);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('24/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Monday',53);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('25/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Tuesday',76);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('26/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Wednesday',96);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('27/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Thursday',145);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('28/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Friday',7);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('29/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Saturday',8);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('30/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Sunday',4);
Insert into EXPORT_TABLE (DT,WEEKDAY,ADMNS) values (to_date('31/OCT/2016 00:00:00','DD/MON/YYYY HH24:MI:SS'),'Monday',6);
答案 0 :(得分：0)
您可以使用
MEDIAN作为分析功能。您只能在其OVER子句中指定query_partition_clause。
因此，您无法在分析查询中指定窗口子句，只能将其过滤到此/去年。相反，您可以使用相关的子查询：
Oracle 11g R2架构设置：
CREATE TABLE export_table(
DT DATE,
WEEKDAY VARCHAR2(9) GENERATED ALWAYS AS ( CAST( TRIM( TO_CHAR( dt, 'Day' ) ) AS VARCHAR2(9) ) ) VIRTUAL,
ADMNS NUMBER
);
Insert into EXPORT_TABLE (DT,ADMNS)
SELECT DATE '2017-10-01',10 FROM DUAL UNION ALL
SELECT DATE '2017-10-02',12 FROM DUAL UNION ALL
SELECT DATE '2017-10-03',15 FROM DUAL UNION ALL
SELECT DATE '2017-10-04',52 FROM DUAL UNION ALL
SELECT DATE '2017-10-05',23 FROM DUAL UNION ALL
SELECT DATE '2017-10-06',54 FROM DUAL UNION ALL
SELECT DATE '2017-10-07',78 FROM DUAL UNION ALL
SELECT DATE '2017-10-08',66 FROM DUAL UNION ALL
SELECT DATE '2017-10-09',5 FROM DUAL UNION ALL
SELECT DATE '2017-10-10',4 FROM DUAL UNION ALL
SELECT DATE '2017-10-11',78 FROM DUAL UNION ALL
SELECT DATE '2017-10-12',99 FROM DUAL UNION ALL
SELECT DATE '2017-10-13',63 FROM DUAL UNION ALL
SELECT DATE '2017-10-14',71 FROM DUAL UNION ALL
SELECT DATE '2017-10-15',63 FROM DUAL UNION ALL
SELECT DATE '2017-10-16',89 FROM DUAL UNION ALL
SELECT DATE '2017-10-17',23 FROM DUAL UNION ALL
SELECT DATE '2017-10-18',45 FROM DUAL UNION ALL
SELECT DATE '2017-10-19',87 FROM DUAL UNION ALL
SELECT DATE '2017-10-20',23 FROM DUAL UNION ALL
SELECT DATE '2017-10-21',12 FROM DUAL UNION ALL
SELECT DATE '2017-10-22',11 FROM DUAL UNION ALL
SELECT DATE '2017-10-23',63 FROM DUAL UNION ALL
SELECT DATE '2017-10-24',93 FROM DUAL UNION ALL
SELECT DATE '2017-10-25',23 FROM DUAL UNION ALL
SELECT DATE '2017-10-26',88 FROM DUAL UNION ALL
SELECT DATE '2017-10-27',44 FROM DUAL UNION ALL
SELECT DATE '2017-10-28',441 FROM DUAL UNION ALL
SELECT DATE '2017-10-29',66 FROM DUAL UNION ALL
SELECT DATE '2017-10-30',23 FROM DUAL UNION ALL
SELECT DATE '2017-10-31',16 FROM DUAL UNION ALL
SELECT DATE '2016-10-01',55 FROM DUAL UNION ALL
SELECT DATE '2016-10-02',67 FROM DUAL UNION ALL
SELECT DATE '2016-10-03',41 FROM DUAL UNION ALL
SELECT DATE '2016-10-04',85 FROM DUAL UNION ALL
SELECT DATE '2016-10-05',25 FROM DUAL UNION ALL
SELECT DATE '2016-10-06',63 FROM DUAL UNION ALL
SELECT DATE '2016-10-07',41 FROM DUAL UNION ALL
SELECT DATE '2016-10-08',87 FROM DUAL UNION ALL
SELECT DATE '2016-10-09',46 FROM DUAL UNION ALL
SELECT DATE '2016-10-10',14 FROM DUAL UNION ALL
SELECT DATE '2016-10-11',74 FROM DUAL UNION ALL
SELECT DATE '2016-10-12',36 FROM DUAL UNION ALL
SELECT DATE '2016-10-13',74 FROM DUAL UNION ALL
SELECT DATE '2016-10-14',63 FROM DUAL UNION ALL
SELECT DATE '2016-10-15',14 FROM DUAL UNION ALL
SELECT DATE '2016-10-16',96 FROM DUAL UNION ALL
SELECT DATE '2016-10-17',31 FROM DUAL UNION ALL
SELECT DATE '2016-10-18',41 FROM DUAL UNION ALL
SELECT DATE '2016-10-19',23 FROM DUAL UNION ALL
SELECT DATE '2016-10-20',96 FROM DUAL UNION ALL
SELECT DATE '2016-10-21',34 FROM DUAL UNION ALL
SELECT DATE '2016-10-22',74 FROM DUAL UNION ALL
SELECT DATE '2016-10-23',63 FROM DUAL UNION ALL
SELECT DATE '2016-10-24',53 FROM DUAL UNION ALL
SELECT DATE '2016-10-25',76 FROM DUAL UNION ALL
SELECT DATE '2016-10-26',96 FROM DUAL UNION ALL
SELECT DATE '2016-10-27',145 FROM DUAL UNION ALL
SELECT DATE '2016-10-28',7 FROM DUAL UNION ALL
SELECT DATE '2016-10-29',8 FROM DUAL UNION ALL
SELECT DATE '2016-10-30',4 FROM DUAL UNION ALL
SELECT DATE '2016-10-31',6 FROM DUAL;
查询1 ：
select dt,
weekday,
admns,
( SELECT MEDIAN( admns )
FROM export_table t
WHERE e.weekday = t.weekday
AND t.dt BETWEEN TRUNC( e.dt, 'MM' )
AND LAST_DAY( e.dt )
) AS this_years_median_for_month,
( SELECT MEDIAN( admns )
FROM export_table t
WHERE e.weekday = t.weekday
AND t.dt BETWEEN ADD_MONTHS( TRUNC( e.dt, 'MM' ), -12 )
AND ADD_MONTHS( LAST_DAY( e.dt ), -12 )
) AS last_years_median_for_month
from export_table e
where EXTRACT( MONTH FROM dt ) = EXTRACT( MONTH FROM SYSDATE )
ORDER BY dt
<强> Results ：
| DT | WEEKDAY | ADMNS | THIS_YEARS_MEDIAN_FOR_MONTH | LAST_YEARS_MEDIAN_FOR_MONTH |
|----------------------|-----------|-------|-----------------------------|-----------------------------|
| 2016-10-01T00:00:00Z | Saturday | 55 | 55 | (null) |
| 2016-10-02T00:00:00Z | Sunday | 67 | 63 | (null) |
| 2016-10-03T00:00:00Z | Monday | 41 | 31 | (null) |
| 2016-10-04T00:00:00Z | Tuesday | 85 | 75 | (null) |
| 2016-10-05T00:00:00Z | Wednesday | 25 | 30.5 | (null) |
| 2016-10-06T00:00:00Z | Thursday | 63 | 85 | (null) |
| 2016-10-07T00:00:00Z | Friday | 41 | 37.5 | (null) |
| 2016-10-08T00:00:00Z | Saturday | 87 | 55 | (null) |
| 2016-10-09T00:00:00Z | Sunday | 46 | 63 | (null) |
| 2016-10-10T00:00:00Z | Monday | 14 | 31 | (null) |
| 2016-10-11T00:00:00Z | Tuesday | 74 | 75 | (null) |
| 2016-10-12T00:00:00Z | Wednesday | 36 | 30.5 | (null) |
| 2016-10-13T00:00:00Z | Thursday | 74 | 85 | (null) |
| 2016-10-14T00:00:00Z | Friday | 63 | 37.5 | (null) |
| 2016-10-15T00:00:00Z | Saturday | 14 | 55 | (null) |
| 2016-10-16T00:00:00Z | Sunday | 96 | 63 | (null) |
| 2016-10-17T00:00:00Z | Monday | 31 | 31 | (null) |
| 2016-10-18T00:00:00Z | Tuesday | 41 | 75 | (null) |
| 2016-10-19T00:00:00Z | Wednesday | 23 | 30.5 | (null) |
| 2016-10-20T00:00:00Z | Thursday | 96 | 85 | (null) |
| 2016-10-21T00:00:00Z | Friday | 34 | 37.5 | (null) |
| 2016-10-22T00:00:00Z | Saturday | 74 | 55 | (null) |
| 2016-10-23T00:00:00Z | Sunday | 63 | 63 | (null) |
| 2016-10-24T00:00:00Z | Monday | 53 | 31 | (null) |
| 2016-10-25T00:00:00Z | Tuesday | 76 | 75 | (null) |
| 2016-10-26T00:00:00Z | Wednesday | 96 | 30.5 | (null) |
| 2016-10-27T00:00:00Z | Thursday | 145 | 85 | (null) |
| 2016-10-28T00:00:00Z | Friday | 7 | 37.5 | (null) |
| 2016-10-29T00:00:00Z | Saturday | 8 | 55 | (null) |
| 2016-10-30T00:00:00Z | Sunday | 4 | 63 | (null) |
| 2016-10-31T00:00:00Z | Monday | 6 | 31 | (null) |
| 2017-10-01T00:00:00Z | Sunday | 10 | 63 | 63 |
| 2017-10-02T00:00:00Z | Monday | 12 | 23 | 31 |
| 2017-10-03T00:00:00Z | Tuesday | 15 | 16 | 75 |
| 2017-10-04T00:00:00Z | Wednesday | 52 | 48.5 | 30.5 |
| 2017-10-05T00:00:00Z | Thursday | 23 | 87.5 | 85 |
| 2017-10-06T00:00:00Z | Friday | 54 | 49 | 37.5 |
| 2017-10-07T00:00:00Z | Saturday | 78 | 74.5 | 55 |
| 2017-10-08T00:00:00Z | Sunday | 66 | 63 | 63 |
| 2017-10-09T00:00:00Z | Monday | 5 | 23 | 31 |
| 2017-10-10T00:00:00Z | Tuesday | 4 | 16 | 75 |
| 2017-10-11T00:00:00Z | Wednesday | 78 | 48.5 | 30.5 |
| 2017-10-12T00:00:00Z | Thursday | 99 | 87.5 | 85 |
| 2017-10-13T00:00:00Z | Friday | 63 | 49 | 37.5 |
| 2017-10-14T00:00:00Z | Saturday | 71 | 74.5 | 55 |
| 2017-10-15T00:00:00Z | Sunday | 63 | 63 | 63 |
| 2017-10-16T00:00:00Z | Monday | 89 | 23 | 31 |
| 2017-10-17T00:00:00Z | Tuesday | 23 | 16 | 75 |
| 2017-10-18T00:00:00Z | Wednesday | 45 | 48.5 | 30.5 |
| 2017-10-19T00:00:00Z | Thursday | 87 | 87.5 | 85 |
| 2017-10-20T00:00:00Z | Friday | 23 | 49 | 37.5 |
| 2017-10-21T00:00:00Z | Saturday | 12 | 74.5 | 55 |
| 2017-10-22T00:00:00Z | Sunday | 11 | 63 | 63 |
| 2017-10-23T00:00:00Z | Monday | 63 | 23 | 31 |
| 2017-10-24T00:00:00Z | Tuesday | 93 | 16 | 75 |
| 2017-10-25T00:00:00Z | Wednesday | 23 | 48.5 | 30.5 |
| 2017-10-26T00:00:00Z | Thursday | 88 | 87.5 | 85 |
| 2017-10-27T00:00:00Z | Friday | 44 | 49 | 37.5 |
| 2017-10-28T00:00:00Z | Saturday | 441 | 74.5 | 55 |
| 2017-10-29T00:00:00Z | Sunday | 66 | 63 | 63 |
| 2017-10-30T00:00:00Z | Monday | 23 | 23 | 31 |
| 2017-10-31T00:00:00Z | Tuesday | 16 | 16 | 75 |
答案 1 :(得分：0)
您可以使用滞后函数，而不是进行两次查询，例如
with a as (select dt, weekday, admns, extract(year from dt) year, to_char(trunc(sysdate,'MM'),'Month') month from export_table),
b as (select dt, weekday, admns, year, month, median(admns) over (partition by year, weekday) med from a)
select dt, weekday, admns, year, month, med, lag(med, extract(day from last_day(dt)) ) over (order by dt) med_prev
from b
order by dt;
这只是使用extract(day from last_day(dt))来获取偏移量，因为样本数据集中只有两个月。
对于全年数据，您可以使用类似
lag(med, dt - add_months(dt,-12))
获取偏移量。 虽然这可能会使用2月29日之前的2月28日中位数，所以如果你不想那样你需要以不同的方式处理这个案例
这也假设数据集包含每天的数据。如果存在可能的差距，则可以使用稍微修改的a cte来填充这些差距以填充感兴趣的时段中的所有日期。