I have this MDX Query:
WITH
MEMBER [COUNT_RANK] AS
RANK(([TKT].[SP].CURRENTMEMBER,
[TKT].[SA].CURRENTMEMBER) ,
[TKT].[SP].CURRENTMEMBER
*[TKT].[SA].[SA]
)
SELECT {
[COUNT_RANK],
[Measures].[TKT Count],
[Measures].[Est Hours]
} ON 0,
ORDER ({[TKT].[SP].[SP]} * {[TKT].[SA].[SA]}, [Measures].[TKT Count], DESC)
ON 1
FROM Ops
the issue I have is that While the COUNT_RANK works and provides a 1-to-n value of ranking per SP for each SA, I need the order of the rank based on TKT Count desc. Meaning for rank = 1, then that SP*SA must have the highest number of TKTs.
Right now the result is random TKT Counts for the RANKING. how do I make the RANK go based on TKT Count DESC?
This is for SQL Server 2016 SSAS. Thanks.
From the MSDN Docs, you should create an ordered set before you declare the RANK() member, and use RANK function on the ordered set.
Here's their example:
WITH
SET OrderedCities AS Order
([Geography].[City].[City].members
, [Measures].[Reseller Sales Amount], BDESC
)
MEMBER [Measures].[City Rank] AS Rank
([Geography].[City].CurrentMember, OrderedCities)
SELECT {[Measures].[City Rank],[Measures].[Reseller Sales Amount]} ON 0
,Order
([Geography].[City].[City].MEMBERS
,[City Rank], ASC)
ON 1
FROM [Adventure Works]
Related
I have a table 'item_queue' containing, items, groups and a sequence number.
Each item is unique and is held against a group with a number indicating the sequence. The count is a total for that item e.g.
group_id|item_id|sequence_order_number|count
--------------------------------------------
A |123 |1 |20
A |124 |2 |30
B |125 |1 |10
Given this information I am trying to set up sequential start and end dates
The start datetime of the first item for a group is the current time, for example assume start of item 123 is '2019-04-04 12:00:00.000' then
end datetime would be start + (count * minutes) so '2019-04-04 12:20:00.000'
The start of item 124 would equal that end date as it is the next in the sequence for that group. the end is then calculated the same way to be '2019-04-04 12:50:00.000'
item 125 would start the time again at '2019-04-04 12:00:00.000' as it is in a different group
I have attempted a few ways to do this, and I think the answer is a recursive cte, but I can't wrap my head around it to make it work for one or multiple groups, my unsuccessful attempt for a single group:
;with cte as
(
select
group_id,
item_id,
count,
GETDATE() as start_datetime,
DATEADD(MINUTE, count, GETDATE()) as end_datetime,
iq.sequence_order_number
from item_queue iq
where iq.group_id = 'A'
union all
select
group_id,
item_id,
count,
cte.end_datetime,
DATEADD(MINUTE, count, cte2.end_datetime) as end_datetime,
iq.sequence_order_number
from item_queue iq
inner join cte
on cte.group_id = iq.group_id
and cte.sequence_order_number > iq.sequence_order_number
where iq.group_id = 'A'
)
select * from cte
I suspect the answer may involve a row number window something like
ROW_NUMBER() OVER (Partition By iq.group_id Order By iq.sequence_order_number ASC)
But I have had trouble using it recursively.
Using SQL server 2012, without the ability to upgrade this database.
The minutes you want to add are practically a cumulative sum. The sum() over() window function is available in 2012 and performs exactly that. Try:
select
*,
isnull(sum([count]) over
(
partition by group_id
order by item_id asc
rows between unbounded PRECEDING and 1 PRECEDING
)
,0) as cum_count_start,
sum([count]) over ( partition by group_id order by item_id asc ) as cum_count_end
from item_queue
You already know how to use dateadd after this point.
What the individual window function caluses do:
partition by group_id : Seperate (partition) the calculations for each group_id value subset
order by item_id asc : make a virtual sorting of the rows on which the window range will be applied
rows between.... : The actual window. For the start date, we want to consider all the lines from the start (thus unbounded preceding) to the previous one (thus 1 preceding), since you don't want the start date to include the current line's [count]. Note that ommitting this clause like we did on the cum_count_end is equivelant to rows between unbounded preceding and current row.
The isnull(...,0) is needed because for the first line of each group_id you want to add 0 to the start date, but the window function sees no rows and returns NULL, so we need to change this to 0.
I am trying to find a way to get the last date by location and product a sum was positive. The only way i can think to do it is with a cursor, and if that's the case I may as well just do it in code. Before i go down that route, i was hoping someone may have a better idea?
Table:
Product, Date, Location, Quantity
The scenario is; I find the quantity by location and product at a particular date, if it is negative i need to get the sum and date when the group was last positive.
select
Product,
Location,
SUM(Quantity) Qty,
SUM(Value) Value
from
ProductTransactions PT
where
Date <= #AsAtDate
group by
Product,
Location
i am looking for the last date where the sum of the transactions previous to and including it are positive
Based on your revised question and your comment, here another solution I hope answers your question.
select Product, Location, max(Date) as Date
from (
select a.Product, a.Location, a.Date from ProductTransactions as a
join ProductTransactions as b
on a.Product = b.Product and a.Location = b.Location
where b.Date <= a.Date
group by a.Product, a.Location, a.Date
having sum(b.Value) >= 0
) as T
group by Product, Location
The subquery (table T) produces a list of {product, location, date} rows for which the sum of the values prior (and inclusive) is positive. From that set, we select the last date for each {product, location} pair.
This can be done in a set based way using windowed aggregates in order to construct the running total. Depending on the number of rows in the table this could be a bit slow but you can't really limit the time range going backwards as the last positive date is an unknown quantity.
I've used a CTE for convenience to construct the aggregated data set but converting that to a temp table should be faster. (CTEs get executed each time they are called whereas a temp table will only execute once.)
The basic theory is to construct the running totals for all of the previous days using the OVER clause to partition and order the SUM aggregates. This data set is then used and filtered to the expected date. When a row in that table has a quantity less than zero it is joined back to the aggregate data set for all previous days for that product and location where the quantity was greater than zero.
Since this may return multiple positive date rows the ROW_NUMBER() function is used to order the rows based on the date of the positive quantity day. This is done in descending order so that row number 1 is the most recent positive day. It isn't possible to use a simple MIN() here because the MIN([Date]) may not correspond to the MIN(Quantity).
WITH x AS (
SELECT [Date],
Product,
[Location],
SUM(Quantity) OVER (PARTITION BY Product, [Location] ORDER BY [Date] ASC) AS Quantity,
SUM([Value]) OVER(PARTITION BY Product, [Location] ORDER BY [Date] ASC) AS [Value]
FROM ProductTransactions
WHERE [Date] <= #AsAtDate
)
SELECT [Date], Product, [Location], Quantity, [Value], Positive_date, Positive_date_quantity
FROM (
SELECT x1.[Date], x1.Product, x1.[Location], x1.Quantity, x1.[Value],
x2.[Date] AS Positive_date, x2.[Quantity] AS Positive_date_quantity,
ROW_NUMBER() OVER (PARTITION BY x1.Product, x1.[Location] ORDER BY x2.[Date] DESC) AS Positive_date_row
FROM x AS x1
LEFT JOIN x AS x2 ON x1.Product=x2.Product AND x1.[Location]=x2.[Location]
AND x2.[Date]<x1.[Date] AND x1.Quantity<0 AND x2.Quantity>0
WHERE x1.[Date] = #AsAtDate
) AS y
WHERE Positive_date_row=1
Do you mean that you want to get the last date of positive quantity come to positive in group?
For example, If you are using SQL Server 2012+:
In following scenario, when the date going to 01/03/2017 the summary of quantity come to 1(-10+5+6).
Is it possible the quantity of following date come to negative again?
;WITH tb(Product, Location,[Date],Quantity) AS(
SELECT 'A','B',CONVERT(DATETIME,'01/01/2017'),-10 UNION ALL
SELECT 'A','B','01/02/2017',5 UNION ALL
SELECT 'A','B','01/03/2017',6 UNION ALL
SELECT 'A','B','01/04/2017',2
)
SELECT t.Product,t.Location,SUM(t.Quantity) AS Qty,MIN(CASE WHEN t.CurrentSum>0 THEN t.Date ELSE NULL END ) AS LastPositiveDate
FROM (
SELECT *,SUM(tb.Quantity)OVER(ORDER BY [Date]) AS CurrentSum FROM tb
) AS t GROUP BY t.Product,t.Location
Product Location Qty LastPositiveDate
------- -------- ----------- -----------------------
A B 3 2017-01-03 00:00:00.000
I am trying to create a column in SQL SERVER that shows 1 OR 0(zero). I have a column of customer numbers that appear more than once. At the first hit on a unique non-repeated customer number it should show one and if it is repeated then 0(zero). How can I create this ?
CustNumber Unique
25122134 1
25122134 0
25122134 0
25122136 1
25122136 0
the solutions I am considering and trying out now are Rank() and Rank_DENSE().
declare #test table
(
CustNumber int
)
insert into #test values
(25122134),
(25122134),
(25122134),
(25122136),
(25122136)
select
* ,
// each CustNumber in partition has the same rank, but different row_number
case when (row_number() over (partition by CustNumber order by CustNumber)) = 1
then 1 else 0 end as [Unique]
// the 1st is unique, the rest (2..n) are not
from #test
order by CustNumber, [Unique] desc
// unique in each group should be displayed first
You don't want RANK because that, by definition, produces the same output for identical inputs.
ROW_NUMBER() and a simple CASE expression should do it:
;WITH Numbered as (
SELECT CustNumber,
ROW_NUMBER() OVER (PARTITION BY CustNumber
ORDER BY CustNumber) as rn --Unusual - pick a real column if you have a preference
FROM YourUnnamedTable
)
SELECT CustNumber,CASE WHEN rn = 1 THEN 1 ELSE 0 END as [Unique]
FROM Numbered
I need a sql server result set that "breaks" on a column value, but if I order by this column in a ranking function, the order I really need is lost. This is best explained by example. The query I'm currently experimenting with is:
select RANK() over(partition by Symbol, Period order by TradeDate desc)
SymbSmaOverUnderGroup, Symbol, TradeDate, Period, Value, Low, LowMinusVal,
LMVSign
from #smasAndLow3
and it returns:
Rnk Symbol TradeDate Period Value Low LowMinusVal LMVSign
1 A 9/6/12 5 37.09 36.71 -.38 U
2 A 9/5/12 5 37.03 36.62 -.41 U
3 A 9/4/12 5 37.07 36.71 -.36 U
4 A 8/31/12 5 37.15 37.30 .15 O
5 A 8/30/12 5 37.22 37.40 .18 O
6 A 8/29/12 5 37.00 36.00 -1.00 U
7 A 8/28/12 5 37.10 37.00 -.10 U
The rank I need here is: 1,1,1,2,2,3,3. So I need to partition by Symbol, Period, and I need to start a new partition on LMVSign (which only contains the values U, O, and E), but it's essential that I order by TradeDate desc. Unless I'm mistaken, partitioning or ordering by LMVSign will make it impossible to sort on the date column. I hope this makes sense. I'm working like mad to do this without a cursor, but I can't get it to work.. thanks in advance.
UPDATE after clarification: I think that you are entering the world of islands and gaps. If your requirement is to group rows by Symbol, Period and LMVSign ordered descendingly by TradeDate, ranking them when any one of these columns change, you might use this (by Itzik Ben-Gan's solution to islands and gaps).
; with islandsAndGaps as
(
select *,
-- Create groups. Important part is order by
-- The difference remains the same as two sequences
-- run along, but the number itself is not ordered
row_number() over (partition by Symbol, Period
order by TradeDate)
- row_number() over (partition by Symbol, Period
order by LMVSign, TradeDate) grp
from Table1
),
grouped as
(
select *,
-- So to order it we use last date in group
-- (mind partition by uses changed order by from second row_number
-- and unordered group number
max(TradeDate) over(partition by LMVSign, grp) DateGroup
from islandsAndGaps
)
-- now we can get rank
select dense_rank() over (order by DateGroup desc) Rnk,
*
from grouped
order by TradeDate desc
Take a look at Sql Fiddle.
OLD answer:
Partition by restarts ranking. I think that you need order by:
dense_rank() over (order by Symbol, Period, LMVSign desc) Rnk
and then you should use TradeDate in order by:
order by Rnk, TradeDate desc
If you need it as a number, add another column:
row_number() over (order by Symbol, Period, LMVSign desc, TradeDate desc) rn
How can I get SQL Server to return the first value (any one, I don't care, it just needs to be fast) it comes across when aggregating?
For example, let's say I have:
ID Group
1 A
2 A
3 A
4 B
5 B
and I need to get any one of the ID's for each group. I can do this as follows:
Select
max(id)
,group
from Table
group by group
which returns
ID Group
3 A
5 B
That does the job, but it seems stupid to me to ask SQL Server to calculate the highest ID when all it really needs to do is to pick the first ID it comes across.
Thanks
PS - the fields are indexed, so maybe it doesn't really make a difference?
There is an undocumented aggregate called ANY which is not valid syntax but is possible to get to appear in your execution plans. This does not provide any performance advantage however.
Assuming the following table and index structure
CREATE TABLE T
(
id int identity primary key,
[group] char(1)
)
CREATE NONCLUSTERED INDEX ix ON T([group])
INSERT INTO T
SELECT TOP 1000000 CHAR( 65 + ROW_NUMBER() OVER (ORDER BY ##SPID) % 3)
FROM sys.all_objects o1, sys.all_objects o2, sys.all_objects o3
I have also populated with sample data such that there are many rows per group.
Your original query
SELECT MAX(id),
[group]
FROM T
GROUP BY [group]
Gives Table 'T'. Scan count 1, logical reads 1367 and the plan
|--Stream Aggregate(GROUP BY:([[T].[group]) DEFINE:([Expr1003]=MAX([[T].[id])))
|--Index Scan(OBJECT:([[T].[ix]), ORDERED FORWARD)
Rewritten to get the ANY aggregate...
;WITH cte AS
(
SELECT *,
ROW_NUMBER() OVER (PARTITION BY [group] ORDER BY [group] ) AS RN
FROM T)
SELECT id,
[group]
FROM cte
WHERE RN=1
Gives Table 'T'. Scan count 1, logical reads 1367 and the plan
|--Stream Aggregate(GROUP BY:([[T].[group]) DEFINE:([[T].[id]=ANY([[T].[id])))
|--Index Scan(OBJECT:([[T].[ix]), ORDERED FORWARD)
Even though potentially SQL Server could stop processing the group as soon as the first value is found and skip to the next one it doesn't. It still processes all rows and the logical reads are the same.
For this particular example with many rows in the group a more efficient version would be a recursive CTE.
WITH RecursiveCTE
AS (
SELECT TOP 1 id, [group]
FROM T
ORDER BY [group]
UNION ALL
SELECT R.id, R.[group]
FROM (
SELECT T.*,
rn = ROW_NUMBER() OVER (ORDER BY (SELECT 0))
FROM T
JOIN RecursiveCTE R
ON R.[group] < T.[group]
) R
WHERE R.rn = 1
)
SELECT *
FROM RecursiveCTE
OPTION (MAXRECURSION 0);
Which gives
Table 'Worktable'. Scan count 2, logical reads 19
Table 'T'. Scan count 4, logical reads 12
The logical reads are much less as it retrieves the first row per group then seeks into the next group rather than reading a load of records that don't contribute to the final result.