Related
I have a T-SQL Quotes table and need to be able to count how many quotes were in an open status during past months.
The dates I have to work with are an 'Add_Date' timestamp and an 'Update_Date' timestamp. Once a quote is put into a 'Closed_Status' of '1' it can no longer be updated. Therefore, the 'Update_Date' effectively becomes the Closed_Status timestamp.
I'm stuck because I can't figure out how to select all open quotes that were open in a particular month.
Here's a few example records:
Quote_No Add_Date Update_Date Open_Status Closed_Status
001 01-01-2016 NULL 1 0
002 01-01-2016 3-1-2016 0 1
003 01-01-2016 4-1-2016 0 1
The desired result would be:
Year Month Open_Quote_Count
2016 01 3
2016 02 3
2016 03 2
2016 04 1
I've hit a mental wall on this one, I've tried to do some case when filtering but I just can't seem to figure this puzzle out. Ideally I wouldn't be hard-coding in dates because this spans years and I don't want to maintain this once written.
Thank you in advance for your help.
You are doing this by month. So, three options come to mind:
A list of all months using left join.
A recursive CTE.
A number table.
Let me show the last:
with n as (
select row_number() over (order by (select null)) - 1 as n
from master..spt_values
)
select format(dateadd(month, n.n, q.add_date), 'yyyy-MM') as yyyymm,
count(*) as Open_Quote_Count
from quotes q join
n
on (closed_status = 1 and dateadd(month, n.n, q.add_date) <= q.update_date) or
(closed_status = 0 and dateadd(month, n.n, q.add_date) <= getdate())
group by format(dateadd(month, n.n, q.add_date), 'yyyy-MM')
order by yyyymm;
This does assume that each month has at least one open record. That seems reasonable for this purpose.
You can use datepart to extract parts of a date, so something like:
select datepart(year, add_date) as 'year',
datepart(month, date_date) as 'month',
count(1)
from theTable
where open_status = 1
group by datepart(year, add_date), datepart(month, date_date)
Note: this counts for the starting month and primarily to show the use of datepart.
Updated as misunderstood the initial request.
Consider following test data:
DECLARE #test TABLE
(
Quote_No VARCHAR(3),
Add_Date DATE,
Update_Date DATE,
Open_Status INT,
Closed_Status INT
)
INSERT INTO #test (Quote_No, Add_Date, Update_Date, Open_Status, Closed_Status)
VALUES ('001', '20160101', NULL, 1, 0)
, ('002', '20160101', '20160301', 0, 1)
, ('003', '20160101', '20160401', 0, 1)
Here is a recursive solution, that doesn't rely on system tables BUT also performs poorer. As we are talking about months and year combinations, the number of recursions will not get overhand.
;WITH YearMonths AS
(
SELECT YEAR(MIN(Add_Date)) AS [Year]
, MONTH(MIN(Add_Date)) AS [Month]
, MIN(Add_Date) AS YMDate
FROM #test
UNION ALL
SELECT YEAR(DATEADD(MONTH,1,YMDate))
, MONTH(DATEADD(MONTH,1,YMDate))
, DATEADD(MONTH,1,YMDate)
FROM YearMonths
WHERE YMDate <= SYSDATETIME()
)
SELECT [Year]
, [Month]
, COUNT(*) AS Open_Quote_Count
FROM YearMonths ym
INNER JOIN #test t
ON (
[Year] * 100 + [Month] <= CAST(FORMAT(t.Update_Date, 'yyyyMM') AS INT)
AND t.Closed_Status = 1
)
OR (
[Year] * 100 + [Month] <= CAST(FORMAT(SYSDATETIME(), 'yyyyMM') AS INT)
AND t.Closed_Status = 0
)
GROUP BY [Year], [Month]
ORDER BY [Year], [Month]
Statement is longer, also more readable and lists all year/month combinations to date.
Take a look at Date and Time Data Types and Functions for SQL-Server 2008+
and Recursive Queries Using Common Table Expressions
I have an sql table like that:
Id Date Price
1 21.09.09 25
2 31.08.09 16
1 23.09.09 21
2 03.09.09 12
So what I need is to get min and max date for each id and dif in days between them. It is kind of easy. Using SQLlite syntax:
SELECT id,
min(date),
max(date),
julianday(max(date)) - julianday(min(date)) as dif
from table group by id
Then the tricky one: how can I receive the price per day during this difference period. I mean something like this:
ID Date PricePerDay
1 21.09.09 25
1 22.09.09 0
1 23.09.09 21
2 31.08.09 16
2 01.09.09 0
2 02.09.09 0
2 03.09.09 12
I create a cte as you mentioned with calendar but dont know how to get the desired result:
WITH RECURSIVE
cnt(x) AS (
SELECT 0
UNION ALL
SELECT x+1 FROM cnt
LIMIT (SELECT ((julianday('2015-12-31') - julianday('2015-01-01')) + 1)))
SELECT date(julianday('2015-01-01'), '+' || x || ' days') as date FROM cnt
p.s. If it will be in sqllite syntax-would be awesome!
You can use a recursive CTE to calculate all the days between the min date and max date. The rest is just a left join and some logic:
with recursive cte as (
select t.id, min(date) as thedate, max(date) as maxdate
from t
group by id
union all
select cte.id, date(thedate, '+1 day') as thedate, cte.maxdate
from cte
where cte.thedate < cte.maxdate
)
select cte.id, cte.date,
coalesce(t.price, 0) as PricePerDay
from cte left join
t
on cte.id = t.id and cte.thedate = t.date;
One method is using a tally table.
To build a list of dates and join that with the table.
The date stamps in the DD.MM.YY format are first changed to the YYYY-MM-DD date format.
To make it possible to actually use them as a date in the SQL.
At the final select they are formatted back to the DD.MM.YY format.
First some test data:
create table testtable (Id int, [Date] varchar(8), Price int);
insert into testtable (Id,[Date],Price) values (1,'21.09.09',25);
insert into testtable (Id,[Date],Price) values (1,'23.09.09',21);
insert into testtable (Id,[Date],Price) values (2,'31.08.09',16);
insert into testtable (Id,[Date],Price) values (2,'03.09.09',12);
The SQL:
with Digits as (
select 0 as n
union all select 1
union all select 2
union all select 3
union all select 4
union all select 5
union all select 6
union all select 7
union all select 8
union all select 9
),
t as (
select Id,
('20'||substr([Date],7,2)||'-'||substr([Date],4,2)||'-'||substr([Date],1,2)) as [Date],
Price
from testtable
),
Dates as (
select Id, date(MinDate,'+'||(d2.n*10+d1.n)||' days') as [Date]
from (
select Id, min([Date]) as MinDate, max([Date]) as MaxDate
from t
group by Id
) q
join Digits d1
join Digits d2
where date(MinDate,'+'||(d2.n*10+d1.n)||' days') <= MaxDate
)
select d.Id,
(substr(d.[Date],9,2)||'.'||substr(d.[Date],6,2)||'.'||substr(d.[Date],3,2)) as [Date],
coalesce(t.Price,0) as Price
from Dates d
left join t on (d.Id = t.Id and d.[Date] = t.[Date])
order by d.Id, d.[Date];
The recursive SQL below was totally inspired by the excellent answer from Gordon Linoff.
And a recursive SQL is probably more performant for this anyway.
(He should get the 15 points for the accepted answer).
The difference in this version is that the datestamps are first formatted to YYYY-MM-DD.
with t as (
select Id,
('20'||substr([Date],7,2)||'-'||substr([Date],4,2)||'-'||substr([Date],1,2)) as [Date],
Price
from testtable
),
cte as (
select Id, min([Date]) as [Date], max([Date]) as MaxDate from t
group by Id
union all
select Id, date([Date], '+1 day'), MaxDate from cte
where [Date] < MaxDate
)
select cte.Id,
(substr(cte.[Date],9,2)||'.'||substr(cte.[Date],6,2)||'.'||substr(cte.[Date],3,2)) as [Date],
coalesce(t.Price, 0) as PricePerDay
from cte
left join t
on (cte.Id = t.Id and cte.[Date] = t.[Date])
order by cte.Id, cte.[Date];
I am trying to get the query for getting the count of users every hour of day in the table. If the data for that hour is not present, I want to record the hour with count of zero. Also users should be counted only for their first entry. Subsequent entries should be ignored.
Table:
userId creationDate
1 2014-10-08 14:33:20.763
2 2014-10-09 04:24:14.283
3 2014-10-10 18:34:26.260
Desired output:
Date UserCount
2014-10-08 00:00:00.000 1
2014-10-08 01:00:00.000 1
2014-10-08 02:00:00.000 1
2014-10-08 03:00:00.000 0
2014-10-08 04:00:00.000 1
....
.....
2014-10-10 23:00:00.000 1
2014-10-10 00:00:00.000 0
My attempt:
SELECT
CAST(creationDate as date) AS ForDate,
DATEPART(hour, date) AS OnHour,
COUNT(distinct userId) AS Totals
FROM
Table
WHERE
primaryKey = 123
GROUP BY
CAST(creationDate as date), DATEPART(hour, createDate)
This only gives me per hour for the record that is present. Not the data for the missing hours. I think there is a way by using a cross join to get 0 data even for the missing hours.
Something like this, I came across, but not able to construct a proper query with it.
cross join (select
ROW_NUMBER() over (order by (select NULL)) as seqnum
from
INFORMATION_SCHEMA.COLUMNS) hours
where hours.seqnum >= 24
Once again, I am not a SQL expert, but trying hard to construct this result set.
One more attempt :
with dh as (
select DATEADD(hour, seqnum - 1, thedatehour ) as DateHour
from (select distinct cast(cast(createDate as DATE) as datetime) as thedatehour
from Table a
) a
cross join
(select ROW_NUMBER() over (order by (select NULL)) as seqnum
from INFORMATION_SCHEMA.COLUMNS
) hours
where hours.seqnum (less than)= 24
)
select dh.DateHour, COUNT(distinct c.userId)
from dh cross join Table c
--on dh.DateHour = c.createDate
group by dh.DateHour
order by 1
You need to build up a table of possible hours, and then join this to your actual records.
The best way to build up a table of possible hours is to use a recursive common table expression.
Here's how:
-- Example data
DECLARE #users TABLE(UserID INT, creationDate DATETIME)
INSERT #users
( UserID, creationDate )
VALUES ( 1, '2014-10-08 14:33:20.763'),
( 2, '2014-10-09 04:24:14.283'),
( 3, '2014-10-10 18:34:26.260')
;WITH u1st AS ( -- determine the FIRST time the user appears
SELECT UserID, MIN(creationDate) AS creationDate
FROM #users
GROUP BY UserID
), hrs AS ( -- recursive CTE of start hours
SELECT DISTINCT CAST(CAST(creationDate AS DATE) AS DATETIME) AS [StartHour]
FROM #users AS u
UNION ALL
SELECT DATEADD(HOUR, 1, [StartHour]) AS [StartHour] FROM hrs
WHERE DATEPART(HOUR,[StartHour]) < 23
), uGrp AS ( -- your data grouped by start hour
SELECT -- note that DATETIMEFROMPARTS is only in SQL Server 2012 and later
DATETIMEFROMPARTS(YEAR(CreationDate),MONTH(CreationDate),
DAY(creationDate),DATEPART(HOUR, creationDate),0,0,0)
AS StartHour,
COUNT(1) AS UserCount FROM u1st AS u
GROUP BY YEAR(creationDate), MONTH(creationDate), DAY(creationDate),
DATEPART(HOUR, creationDate)
)
SELECT hrs.StartHour, ISNULL(uGrp.UserCount, 0) AS UserCount
FROM hrs LEFT JOIN uGrp ON hrs.StartHour = uGrp.StartHour
ORDER BY hrs.StartHour
NB - DATETIMEFROMPARTS is only in SQL SERVER 2012 and greater. If you are using an earlier version of SQL SERVER you could have
WITH u1st AS ( -- determine the FIRST time the user appears
SELECT UserID, MIN(creationDate) AS creationDate
FROM #users
GROUP BY UserID
), hrs AS ( -- recursive CTE of start hours
SELECT DISTINCT CAST(CAST(creationDate AS DATE) AS DATETIME) AS [StartHour]
FROM #users AS u
UNION ALL
SELECT DATEADD(HOUR, 1, [StartHour]) AS [StartHour] FROM hrs
WHERE DATEPART(HOUR,[StartHour]) < 23
), uGrp AS ( -- your data grouped by start hour
SELECT -- note that DATETIMEFROMPARTS is only in SQL Server 2012 and later
CAST(CAST(YEAR(creationDate) AS CHAR(4)) + '-'
+ RIGHT('0' + CAST(MONTH(creationDate) AS CHAR(2)), 2) + '-'
+ RIGHT('0' + CAST(DAY(creationDate) AS CHAR(2)), 2) + ' '
+ RIGHT('0' + CAST(DATEPART(HOUR, creationDate) AS CHAR(2)), 2)
+ ':00:00.000'
AS DATETIME) AS StartHour,
COUNT(1) AS UserCount FROM u1st AS u
GROUP BY YEAR(creationDate), MONTH(creationDate), DAY(creationDate),
DATEPART(HOUR,creationDate)
)
SELECT hrs.StartHour, ISNULL(uGrp.UserCount, 0) AS UserCount
FROM hrs LEFT JOIN uGrp ON hrs.StartHour = uGrp.StartHour
ORDER BY hrs.StartHour
I asked a similar question on dba just this morning...https://dba.stackexchange.com/questions/86435/filling-in-date-holes-in-grouped-by-date-sql-data.
You can used my GetSequence function, or create a Numbers table. I haven't done my own testing yet to validate what was suggested in my scenario.
Try this:
BUILD SAMPLE DATA
CREATE TABLE yourTable(
userId INT,
creationDate DATETIME
)
INSERT INTO yourTable VALUES (1, '2014-10-08 14:33:20.763'), (2, '2014-10-09 04:24:14.283'),(3, '2014-10-10 18:34:26.260');
SOLUTION
WITH tally(N) AS(
SELECT TOP(23) ROW_NUMBER() OVER(ORDER BY (SELECT NULL)) FROM sys.columns
)
,hourly(creationDate) AS(
SELECT DATEADD(HOUR, t.N, d.creationDate)
FROM tally t
CROSS JOIN(
SELECT DISTINCT DATEADD(DD, DATEDIFF(DD, 0, creationDate), 0) AS creationDate FROM yourTable
) d
)
SELECT
h.creationDate,
userCount = ISNULL(t.userCount, 0)
FROM hourly h
LEFT JOIN(
SELECT
creationDate = DATEADD(HOUR, DATEPART(HOUR, creationDate) ,DATEADD(DD, DATEDIFF(DD, 0, creationDate), 0)),
userCount = COUNT(*)
FROM yourTable
GROUP BY DATEADD(DD, DATEDIFF(DD, 0, creationDate), 0), DATEPART(HOUR, creationDate)
)t
ON t.creationDate = h.creationDate
CLEANUP
DROP TABLE yourTable
Create a temporary table (let's say #CreationDateHours) containing create date and hours from 0 to 23.
Declare #date as date
SELECT MAX(CAST(creationDate as date)) AS ForDate, 0 as OnHour into #CreationDateHours
FROM Table
WHERE
primaryKey = 123
Select #date=ForDate from #CreationDateHours
Declare #i int
Set #i=1
While #i<24
begin
insert into #CreationDateHours
select #date as ForDate, #i as OnHour
set #i+=1
end
Now, Run this query to get the desired results
select t1.ForDate, t1.OnHour, isnull(t2.Totals,0) AS Totals
from
#CreationDateHours t1 left join (SELECT
CAST(creationDate as date) AS ForDate,
DATEPART(hour, date) AS OnHour,
COUNT(distinct userId) AS Totals
FROM
Table
WHERE
primaryKey = 123
GROUP BY
CAST(creationDate as date), DATEPART(hour, createDate)) as t2
on t1.ForDate= t2.ForDate and t1.OnHour=t2.OnHour
select count, strftime('%H', creationDate) as hour from table group by hour;
OUTUPUT:
count hour
n1 01
n2 02
n3 03
... ...
n24 24
Tested on SQLite3
https://www.sqlite.org/lang_datefunc.html
you can see all the formats
(like %d for every day of month)
if you want the whole date
strftime('%Y-%m-%d %H, creationDate)
Suppose I have a table that indicates the number of items sold in a particular month for each sales rep. However, there will not be a row for a particular person in months where there were no sales. Example
rep_id month_yr num_sales
1 01/01/2012 3
1 05/01/2012 1
1 11/01/2012 1
2 02/01/2012 2
2 05/01/2012 1
I want to be able to create a query that shows for each rep_id and all possible months (01/01/2012, 02/01/2012, etc. through current) a rolling 12 month sales sum, like this:
rep_id month_yr R12_Sum
1 11/01/2012 5
1 12/01/2012 5
1 01/01/2013 5
1 02/01/2013 2
I have found some examples online, but the problem I'm running into is I'm missing some dates for each rep_id. Do I need to cross join or something?
To solve this problem, you need a driver table that has all year/month combinations. Then, you need to create this for each rep.
The solution is then to left join the actual data to this driver and aggregate the period that you want. Here is the query:
with months as (
select 1 as mon union all select 2 union all select 3 union all select 4 union all
select 5 as mon union all select 6 union all select 7 union all select 8 union all
select 9 as mon union all select 10 union all select 11 union all select 12
),
years as (select 2010 as yr union all select 2011 union all select 2012 union all select 2013
),
monthyears as (
select yr, mon, yr*12+mon as yrmon
from months cross join years
),
rmy as (
select *
from monthyears my cross join
(select distinct rep_id from t
) r
)
select rmy.rep_id, rmy.yr, rmy.mon, SUM(t.num_sales) as r12_sum
from rmy join
t
on rmy.rep_id = t.rep_id and
t.year(month_yr)*12 + month(month_yr) between rmy.yrmon - 11 and rmy.yrmon
group by rmy.rep_id, rmy.yr, rmy.mon
order by 1, 2, 3
This hasn't been tested, so it may have syntactic errors. Also, it doesn't convert the year/month combination back to a date, leaving the values in separate columns.
Here is one solution:
SELECT
a.rep_id
,a.month_yr
,SUM(b.R12_Sum) AS R12_TTM
FROM YourTable a
LEFT OUTER JOIN YourTable b
ON a.rep_id = b.rep_id
AND a.month_yr <= b.month_yr
AND a.month_yr >= DATEADD(MONTH, -11, b.month_yr)
GROUP BY
a.rep_id
,a.month_yr
It's certainly not pretty but is more simple than a CTE, numbers table or self join:
DECLARE #startdt DATETIME
SET #startdt = '2012-01-01'
SELECT rep_id, YEAR(month_yr), MONTH(month_yr), SUM(num_sales)
FROM MyTable WHERE month_yr >= #startdt AND month_yr < DATEADD(MONTH,1,#startdt)
UNION ALL
SELECT rep_id, YEAR(month_yr), MONTH(month_yr), SUM(num_sales)
FROM MyTable WHERE month_yr >= DATEADD(MONTH,1,#startdt) AND month_yr < DATEADD(MONTH,2,#startdt)
UNION ALL
SELECT rep_id, YEAR(month_yr), MONTH(month_yr), SUM(num_sales)
FROM MyTable WHERE month_yr >= DATEADD(MONTH,2,#startdt) AND month_yr < DATEADD(MONTH,3,#startdt)
UNION ALL
SELECT rep_id, YEAR(month_yr), MONTH(month_yr), SUM(num_sales)
FROM MyTable WHERE month_yr >= DATEADD(MONTH,3,#startdt) AND month_yr < DATEADD(MONTH,4,#startdt)
UNION ALL
etc etc
The following demonstrates using a CTE to generate a table of dates and generating a summary report using the CTE. Sales representatives are omitted from the results when they have had no applicable sales.
Try jiggling the reporting parameters, e.g. setting #RollingMonths to 1, for more entertainment.
-- Sample data.
declare #Sales as Table ( rep_id Int, month_yr Date, num_sales Int );
insert into #Sales ( rep_id, month_yr, num_sales ) values
( 1, '01/01/2012', 3 ),
( 1, '05/01/2012', 1 ),
( 1, '11/01/2012', 1 ),
( 2, '02/01/2012', 1 ),
( 2, '05/01/2012', 2 );
select * from #Sales;
-- Reporting parameters.
declare #ReportEnd as Date = DateAdd( day, 1 - Day( GetDate() ), GetDate() ); -- The first of the current month.
declare #ReportMonths as Int = 6; -- Number of months to report.
declare #RollingMonths as Int = 12; -- Number of months in rolling sums.
-- Report.
-- A CTE generates a table of month/year combinations covering the desired reporting time period.
with ReportingIntervals as (
select DateAdd( month, 1 - #ReportMonths, #ReportEnd ) as ReportingInterval,
DateAdd( month, 1 - #RollingMonths, DateAdd( month, 1 - #ReportMonths, #ReportEnd ) ) as FirstRollingMonth
union all
select DateAdd( month, 1, ReportingInterval ), DateAdd( month, 1, FirstRollingMonth )
from ReportingIntervals
where ReportingInterval < #ReportEnd )
-- Join the CTE with the sample data and summarize.
select RI.ReportingInterval, S.rep_id, Sum( S.num_sales ) as R12_Sum
from ReportingIntervals as RI left outer join
#Sales as S on RI.FirstRollingMonth <= S.month_yr and S.month_yr <= RI.ReportingInterval
group by RI.ReportingInterval, S.rep_id
order by RI.ReportingInterval, S.rep_id
I have a table like:
id month cost
------------------
1 Jan 200
1 Mar 204
1 May 200
1 Dec 201
I need an output like( order by month including the other months of a year-displaying all 12 months):
to month cost
------------------
1 Jan 200
NULL Feb NULL
1 Mar 204
....
....
....
1 Dec 201
any idea or solution how to do this in TSQL?
Thanks!
edit:: month is extracted from a datetime value.
in real world i'll have to show previous 12 months from last month in a DESC order! any suggestion for that?
Try building a reference table of months, and JOINing on it. It's the quickest way to do this with months in varchar datatype.
declare #foo table (id int, [mon] varchar(100), cost int)
declare #mon table (mon varchar(100), orderWeight int)
INSERT INTO #mon (mon, orderWeight)
VALUES ('Jan',1), ('Feb',2),('Mar',3),('Apr',4),('May',5),('Jun',6),('Jul',7),
('Aug',8),('Sep',9),('Oct',10),('Nov',11),('Dec',12)
INSERT INTO #foo(id, [mon], cost)
VALUES ( 1 ,'Jan' , 200),
( 1 ,'Mar', 204),
( 1 ,'May' , 200),
( 1 ,'Dec' , 201)
select f.id,
m.[mon] ,
f.cost
from #mon as m
left join #foo as f on m.mon = f.mon
order by m.orderWeight
Results:
Your ordering will now be guaranteed with the order by orderWeight.
Sample table
create table mytable(id int, dt datetime, cost money)
insert mytable values
(1,GETDATE()-10,200),
(1,GETDATE()-40,204),
(1,GETDATE()-100,200),
(1,GETDATE()-200,201);
The query, using SQL Server 2008 specific syntax, and sorted properly
select
t.id [to],
CONVERT(char(3),dateadd(month,-M.N,L.PVT),7) [Month],
sum(t.cost) totalCost
from (select PVT=dateadd(month,datediff(month,0,getdate())-1,0)) L
cross join (values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9),(10),(11)) M(N)
left join mytable t
on t.dt >= dateadd(month,-M.N,L.PVT)
and t.dt < dateadd(month,-M.N+1,L.PVT)
group by t.id, right(CONVERT(char(9),dt,6),6), M.N, L.PVT
order by M.N
What it does:
right(CONVERT(char(9),dt,6),6) converts a date into the format 'DD MMM YY', we only need the MMM YY part
In the SELECT, we further extract only the 3-char month from it, using LEFT( , 3)
The subquery L has a single record and column, PVT, which is the first date of the last month
The number series 0-11 is used to create the month values for the last 12 months, using the formula dateadd(month,-M.N,L.PVT)
The range t.dt >= .. and t.dt < .. finds data for a single month
How about this?
The result contains month and year, but you can strip it as you want.
;with months
as
(
select dateadd(month, -1, dateadd(day, datediff(day, 0, getdate()), 0)) as m
union all
select dateadd(month, -1, m)
from months
where m > dateadd(month, -12, getdate())
)
-- Testdata
,yourTable(id,somedate,cost)
as
(
select 1, '2011-01-03', 200
union all
select 1, '2011-03-06', 204
union all
select 1, '2010-05-09', 200
union all
select 1, '2010-05-19', 201
union all
select 1, '2010-12-02', 201
)
-- end testdata
select yt.id
,datename(month,coalesce(yt.somedate, m.m)) as [month]
,datename(year,coalesce(yt.somedate, m.m)) as [year]
--,yt.cost
,sum(yt.cost) as cost
from months m
left join yourTable yt
on datepart(year, yt.someDate) = DATEPART(year, m.m)
and datepart(month, yt.someDate) = DATEPART(month, m.m)
group by
yt.id
,datename(month,coalesce(yt.somedate, m.m))
,datename(year,coalesce(yt.somedate, m.m))
,m.m
order by m.m desc
Edit: Altered solution to support sum.
Remove the group by-section and alter the comment of cost, to get the old solution.