选择特定时期的前 5 个最大高点和 5 分钟低点，例如在同一查询中过去 180 天和 30 天

Question

Following is the table data, need to find out the top 5 max high and min low for the entire 180 days and the same for the last 30 days in the same query.

Stock  High Low Date        prevclose ....
------------------------------------
ABB    100   75  29/12/2019   90
ABB     83   50  30/12/2019   87
ABB     73   45  30/12/2019   87
.
.
.
.
ABB    100    67  29/06/2019  90
ABB     83    65  30/06/2019  81
infy   100    75  29/12/2019  90
infy    830  650  30/12/2019  810
infy    730  645  30/12/2019  788 
.
.
.
infy   1001  556  29/06/2019  904
infy    833  657  30/06/2019  812
infy    734  643  30/06/2019  735

Query, which I tried, but getting an error at rank() cannot be used with window functions. any alternatives.

select * into SRTREND180 from (
select *
from (
    select 
        rank() over(partition by name order by high desc) rn_high180,
        rank() over(partition by name order by low asc) rn_low180,
        rank() over (partition by name order by high desc rows between 30 preceding and current row) rn_high30,
        rank() over (partition by name order by low asc rows between 30 preceding and current row) rn_low30,        
         t.*
    from Historic t
) Hist
where rn_high180 <= 5 or rn_low180 <= 5 or  rn_high30 <=5 or rn_low30 <=5
) SR

Answer 1

"rows between 30 preceding and current row" is the window function that is causing the error. Try separating out the two select fields for the last 30 days into a separate query, or put them in a sub-query selecting off of the data with the windowed part of the select moved to the where clause.

EDIT: Try using the following CTEs at the start of the query (change the table/field names to yours). This will allow you to write a query while having two separate data tables, one with the ranked lows/highs for the last 180 days, and one for the last 30 days.

with last180 (name, closedate, low, high, lowrank, highrank) as
(
    select name, closedate, low, high, rank() over(partition by name order by low asc), rank() over(partition by name order by high desc)
    from @table where datediff(day, closedate, getdate()) <= 180
),
last30 (name, closedate, low, high, lowrank, highrank) as
(
    select name, closedate, low, high, rank() over(partition by name order by low asc), rank() over(partition by name order by high desc)
    from @table where datediff(day, closedate, getdate()) <= 30
)

Answer 2

One workaround to rank not working on records in partition is to add a subquery to virtualize the buckets and then use the bucket marker as part of the partition as needed.

SQL Fiddle

MS SQL Server 2017 Schema Setup :

CREATE TABLE T (name NVARCHAR(20), High INT, Low INT, Date DATETIME, PrevClose INT)
INSERT T VALUES
('ABB', 100, 75,'12/29/2019',90),
('ABB', 83,  50,'12/30/2019',87),
('ABB', 73, 45,'12/30/2019',87),
('ABB', 100, 67,'06/29/2019',90),
('ABB', 83, 65,'06/30/2019',81),
('INFY', 100, 75,'12/29/2019',90),
('INFY', 830, 600,'12/30/2019',810),
('INFY', 730, 645,'12/30/2019',788),
('INFY', 1001, 556,'06/29/2019',904),
('INFY', 833, 657,'06/30/2019',812),
('INFY', 734,643, '06/30/2019',735),
('INFY', 734,643, '07/30/2019',735)

Query 1 :

DECLARE @ReportDate DATETIME = GETDATE()

;WITH DataWithDayFlag AS
(
    select 
        *,
        DaysOut   =  DATEDIFF(DAY,date,@ReportDate),
        Bucket30  = CASE WHEN DATEDIFF(DAY,date,@ReportDate) <= 30 THEN 1 ELSE NULL END,
        Bucket180 = CASE WHEN DATEDIFF(DAY,date,@ReportDate) <= 180 THEN 1 ELSE NULL END

    FROM
        T
)
SELECT
  CASE WHEN Bucket180 IS NOT NULL THEN rank() over (partition by name, Bucket180 order by high desc) ELSE NULL END rn_high180,
  CASE WHEN Bucket180 IS NOT NULL THEN rank() over (partition by name, Bucket180 order by low asc) ELSE NULL END rn_low180,
  CASE WHEN Bucket30 IS NOT NULL THEN rank() over (partition by name, Bucket30 order by high desc) ELSE NULL END rn_high30,
  CASE WHEN Bucket30 IS NOT NULL THEN rank() over (partition by name, Bucket30 order by low asc) ELSE NULL END rn_low30,        
  t.*
from 
  DataWithDayFlag t
where  
  DaysOut <= 180
ORDER BY
    name

Results :

| rn_high180 | rn_low180 | rn_high30 | rn_low30 | name | High | Low |                 Date | PrevClose | DaysOut | Bucket30 | Bucket180 |
|------------|-----------|-----------|----------|------|------|-----|----------------------|-----------|---------|----------|-----------|
|          3 |         1 |         3 |        1 |  ABB |   73 |  45 | 2019-12-30T00:00:00Z |        87 |       1 |        1 |         1 |
|          2 |         2 |         2 |        2 |  ABB |   83 |  50 | 2019-12-30T00:00:00Z |        87 |       1 |        1 |         1 |
|          1 |         3 |         1 |        3 |  ABB |  100 |  75 | 2019-12-29T00:00:00Z |        90 |       2 |        1 |         1 |
|          2 |         3 |    (null) |   (null) | INFY |  734 | 643 | 2019-07-30T00:00:00Z |       735 |     154 |   (null) |         1 |
|          4 |         1 |         3 |        1 | INFY |  100 |  75 | 2019-12-29T00:00:00Z |        90 |       2 |        1 |         1 |
|          1 |         2 |         1 |        2 | INFY |  830 | 600 | 2019-12-30T00:00:00Z |       810 |       1 |        1 |         1 |
|          3 |         4 |         2 |        3 | INFY |  730 | 645 | 2019-12-30T00:00:00Z |       788 |       1 |        1 |         1 |