Group entries per week where each column got specific date validation - sql

Question

I am trying to group tickets per week in a year when data field X are between that week. The final result should be something like Week,Datefield1,Datefield2..
What I reached so far:
DECLARE #YearStartDate datetime
SET #YearStartDate = '2016-01-04 00:00:00.000'
DECLARE #YearEndDate datetime
SET #YearEndDate = '2016-04-24 00:00:00.000'
SELECT
Year(T.CreateDate) as 'Year',
'Week ' + cast(datepart(wk, CreateDate) as varchar(2)) as 'Week',
Sum(CASE WHEN CreateDate BETWEEN #YearStartDate and #YearEndDate
THEN 1 Else 0 End) 'Created'
FROM mytable AS T
GROUP BY Year(T.CreateDate),datepart(wk,T.CreateDate)
ORDER BY Year(T.CreateDate),datepart(wk, T.CreateDate)
Result:
Year Week Created
----------- ------- -----------
2016 Week 1 0
2016 Week 2 5
2016 Week 3 3
2016 Week 4 2
Goal:
Year Week Created Schuduled Closed
----------- ------- ----------- ----------- -----------
2016 Week 1 0 0 0
2016 Week 2 5 3 2
2016 Week 3 3 2 2
2016 Week 4 2 2 0
Suggestion?
SELECT
Year(..) as 'Year',
'Week ' + cast(datepart(wk, ...) as varchar(2)) as 'Week',
Sum(CASE WHEN CreateDate BETWEEN #YearStartDate and #YearEndDate
THEN 1 Else 0 End) 'Created'
Sum(CASE WHEN ScheduledDate BETWEEN #YearStartDate and #YearEndDate
THEN 1 Else 0 End) 'Scheduled'
Sum(CASE WHEN ClosedDate BETWEEN #YearStartDate and #YearEndDate
THEN 1 Else 0 End) 'Closed'
FROM mytable AS T
GROUP BY Year(...)
ORDER BY Year(...)
Constraints:
I can count the entries per week for one datafield but for several datafields (created,scheduled,pending,closed) I think I need to change the logic. Maybe create a inner join or left join against the same table.
I was wondering if you could give me some guidance on finding the right path. Thanks for taking the time to help me.
All the best.

Answer 1

Right, join on other similar SELECTs for scheduled and closed should work. Something as
DECLARE #YearStartDate datetime = '2016-01-04 00:00:00.000';
DECLARE #YearEndDate datetime = '2016-04-24 00:00:00.000';
SELECT CLR.*, SCH.Schuduled, CLS.Closed FROM
(SELECT
Year(T.CreateDate) as 'Year',
'Week ' + cast(datepart(wk, CreateDate) as varchar(2)) as 'Week',
Sum(CASE WHEN CreateDate BETWEEN #YearStartDate and #YearEndDate THEN 1 Else 0 End) 'Created'
FROM mytable AS T
GROUP BY Year(T.CreateDate),datepart(wk,T.CreateDate)
) CRT
JOIN
(SELECT
Year(T.ScheduleDate) as 'Year',
'Week ' + cast(datepart(wk, ScheduleDate) as varchar(2)) as 'Week',
Sum(CASE WHEN ScheduleDate BETWEEN #YearStartDate and #YearEndDate THEN 1 Else 0 End) 'Schuduled'
FROM mytable AS T
GROUP BY Year(T.ScheduleDate),datepart(wk,T.ScheduleDate)
) SCH
ON CRT.Year = SCH.Year
AND CRT.Week = SCH.Week
JOIN
(SELECT
Year(T.ClosedDate) as 'Year',
'Week ' + cast(datepart(wk, ClosedDate) as varchar(2)) as 'Week',
Sum(CASE WHEN ClosedDate BETWEEN ClosedDate and #YearEndDate THEN 1 Else 0 End) 'Closed'
FROM mytable AS T
GROUP BY Year(T.ClosedDate),datepart(wk,T.ClosedDate)
) CLS
ON CRT.Year = CLS.Year
AND CRT.Week = CLS.Week
ORDER BY Year, Week

Answer 2

You have three different date columns. One method would pre-aggregation along the dimensions and then use full outer join. Alternatively, you can use union all and aggregation:
SELECT Year(dte) as [Year],
'Week ' + datename(wk, dte) as [Week],
Sum(Created) as Created,
Sum(Scheduled) as Scheduled,
Sum(Closed) as Closed
FROM ((SELECT createddate as dte, 1 as created, 0 as schedule, 0 as closed
FROM mytable
) UNION ALL
(SELECT scheduleddate as dte, 0 as created, 1 as schedule, 0 as closed
FROM mytable
) UNION ALL
(SELECT closeddate as date, 1 as created, 0 as schedule, 0 as closed
FROM mytable
)
) t
GROUP BY Year(date), datename(wk, dte)
ORDER BY Year(date), datename(wk, dte);
I would encourage a couple of things:
Do not use single quotes for column aliases. Only use single quotes for string and date names.
Do not use reserved works such as "year" for column or table names.
Also note the use of datename(). This returns a string so no conversion is needed.