This question relates to this post: Split date range into one row per month in sql server
But I'm new to Stackoverflow and new users are not allowed to place comments to provided solutions.That's why I am now adding this as a new question.
The question was this:
I have a table with two column called "from_date" and "to_date"
The table looks like this:
Table (I am not allowed to embed my picture here, so please click the link)
I want result like:-
from_date || to_date
----------- ------------
2013-11-25 || 2013-11-30
2013-12-01 || 2013-12-05
That date is splits from 2013-11-25 to 2013-11-30 and another date split from 2013-12-01 to 2013-12-05... Is it possible to split like this ?
Aaron Bertrand's solution is this:
DECLARE #d TABLE(from_date DATE, to_date DATE);
INSERT #d VALUES ('2013-11-25','2013-12-05');
;WITH n(n) AS
(
SELECT ROW_NUMBER() OVER (ORDER BY [object_id])-1
FROM sys.all_columns
),
d(n,f,t,md,bp,ep) AS
(
SELECT n.n, d.from_date, d.to_date,
DATEDIFF(MONTH, d.from_date, d.to_date),
DATEADD(MONTH, n.n, DATEADD(DAY, 1-DAY(from_date), from_date)),
DATEADD(MONTH, n.n+1, DATEADD(DAY, 0-DAY(from_date), from_date))
FROM n
INNER JOIN #d AS d ON d.to_date >= DATEADD(MONTH, n.n-1, d.from_date)
)
SELECT original_from_date = f, original_to_date = t,
new_from_date = CASE n WHEN 0 THEN f ELSE bp END,
new_to_date = CASE n WHEN md THEN t ELSE ep END
FROM d
WHERE md >= n
ORDER BY original_from_date, new_from_date;
Results:
original_from_date original_to_date new_from_date new_to_date
------------------ ---------------- ------------- -----------
2013-11-25 2013-12-05 2013-11-25 2013-11-30
2013-11-25 2013-12-05 2013-12-01 2013-12-05
Aaron Bertrand's answer is a super powerful query, but it seems it only works well if the startdate is in a month following a month with 31 days. If you change the startdate to, for example, 2013-5-25 (following April, with only 30 days), the new_to_date values are incorrect: all full months suddenly get a max of 30 days then:
original_from_date original_to_date new_from_date new_to_date
2013-05-25 2013-12-05 2013-05-25 2013-05-30
2013-05-25 2013-12-05 2013-06-01 2013-06-30
2013-05-25 2013-12-05 2013-07-01 2013-07-30
2013-05-25 2013-12-05 2013-08-01 2013-08-30
2013-05-25 2013-12-05 2013-09-01 2013-09-30
2013-05-25 2013-12-05 2013-10-01 2013-10-30
2013-05-25 2013-12-05 2013-11-01 2013-11-30
2013-05-25 2013-12-05 2013-12-01 2013-12-05
Can anyone edit the query so that end dates are correct (28,30,31 days, and 29 feb leap) regardless of what month the start date is in? I can't figure out how to do it.
Many thanks, Martijn Vermunt
The answer has been provided, the query has been edited and made perfect. Thank you #Aaron Bertrand
for updating the query!
DECLARE #d TABLE(from_date DATE, to_date DATE);
INSERT #d VALUES ('2013-11-25','2013-12-05');
;WITH n(n) AS
(
SELECT ROW_NUMBER() OVER (ORDER BY [object_id])-1 FROM sys.all_columns
),
d(n,f,t,md,bp,ep) AS
(
SELECT n.n, d.from_date, d.to_date,
DATEDIFF(MONTH, d.from_date, d.to_date),
DATEADD(MONTH, n.n, DATEADD(DAY, 1-DAY(from_date), from_date)),
DATEADD(DAY, -1, DATEADD(MONTH, 1, DATEADD(MONTH, n.n,
DATEADD(DAY, 1-DAY(from_date), from_date))))
FROM n INNER JOIN #d AS d
ON d.to_date >= DATEADD(MONTH, n.n-1, d.from_date)
)
SELECT original_from_date = f, original_to_date = t,
new_from_date = CASE n WHEN 0 THEN f ELSE bp END,
new_to_date = CASE n WHEN md THEN t ELSE ep END
FROM d WHERE md >= n
ORDER BY original_from_date, new_from_date;
Related
i have this ms-sql code below, running on sql server 2017:
DECLARE #report_int TABLE
(
atimestamp DATETIME,
adate DATE,
ahour INT,
intvalue INT,
attribute CHAR(254),
total INT
)
INSERT INTO #report_int
SELECT
el.servertimestamp,
cast(el.servertimestamp as date) adate,
(DATEPART(HH,el.servertimestamp)) ahour,
el.intvalue,
el.attributes,
0
FROM
eventlog el
where
el.servertimestamp BETWEEN '6/29/2017'
AND DATEADD(dd, +1, '6/29/2019')
and (attributes LIKE '%N<=>OlympieioIn%' OR attributes LIKE '%N<=>OlympieioOut%')
order by
el.servertimestamp
DECLARE #report TABLE
(
adate datetime,
ahour int,
came_in bigint,
gone_out bigint
)
--fill report with dates and hours
INSERT INTO #report (adate, ahour) SELECT DISTINCT adate, ahour FROM #report_int
--reset
UPDATE #report set came_in = 0, gone_out = 0
--update in totals, min and max times
UPDATE #report SET came_in = i.total from
#report r,
(
select adate, ahour, COUNT(*) total
FROM #report_int
WHERE attribute LIKE 'N<=>OlympieioIn'
GROUP BY adate, ahour
) i
WHERE
i.adate = r.adate
and i.ahour = r.ahour
--update out totals
UPDATE #report SET gone_out = i.total from
#report r,
(
SELECT adate, ahour, COUNT(*) total
FROM #report_int
WHERE attribute LIKE 'N<=>OlympieioOut'
GROUP BY adate, ahour
) i
where
i.adate = r.adate
and i.ahour = r.ahour
select
adate,
ahour,
RIGHT('00'+cast(ahour as varchar),2)+':00 - '+RIGHT('00'+cast(ahour+1 as varchar),2)+':00' timeframe,
came_in,
gone_out
from
#report
This code shows how many people came in or gone out of a building every day. As you can see, the amount of people who come in and go out, are displayed in an hourly grouping.
Below is a screenshot of the executed code.
What i want to do is group these data in a 2-hour grouping.For example, the grouping of the 3rd and 4th hour of the executed code above, will be:
DATE(adate): 2018-05-12
Hour(ahour): 15
Timeframe: 15:00-17:00
came_in: 0
gone_out: 2
Your help will be much appreciated,
thank you.
One way to get a datetime to the last 2 hour slot would be to use:
DATEADD(HOUR, DATEDIFF(HOUR, 0, YourDateColumn) - (DATEDIFF(HOUR, 0, YourDateColumn) % 2), 0)
If we then take some (random) sample data, you get the following:
WITH VTE AS (
SELECT CONVERT(datetime2(0),V.DT) AS DT
FROM (VALUES('2018-07-04T00:11:32'),('2018-07-04T01:17:12'),('2018-07-04T02:00:01'),('2018-07-04T07:49:59'),('2018-07-04T07:59:58'),('2018-07-04T08:00:00')) V(DT))
SELECT DT, DATEADD(HOUR, DATEDIFF(HOUR, 0, DT) - (DATEDIFF(HOUR, 0, DT) % 2), 0) AS DT2h
FROM VTE;
Which returns:
DT DT2h
--------------------------- -----------------------
2018-07-04 00:11:32 2018-07-04 00:00:00.000
2018-07-04 01:17:12 2018-07-04 00:00:00.000
2018-07-04 02:00:01 2018-07-04 02:00:00.000
2018-07-04 07:49:59 2018-07-04 06:00:00.000
2018-07-04 07:59:58 2018-07-04 06:00:00.000
2018-07-04 08:00:00 2018-07-04 08:00:00.000
If you wanted a COUNT by 2 hour slot:
WITH VTE AS (
SELECT CONVERT(datetime2(0),V.DT) AS DT
FROM (VALUES('2018-07-04T00:11:32'),('2018-07-04T01:17:12'),('2018-07-04T02:00:01'),('2018-07-04T07:49:59'),('2018-07-04T07:59:58'),('2018-07-04T08:00:00')) V(DT))
SELECT DATEADD(HOUR, DATEDIFF(HOUR, 0, DT) - (DATEDIFF(HOUR, 0, DT) % 2), 0) AS DT2h, COUNT(DT) AS C
FROM VTE
GROUP BY DATEADD(HOUR, DATEDIFF(HOUR, 0, DT) - (DATEDIFF(HOUR, 0, DT) % 2), 0);
Which returns:
DT2h C
----------------------- -----------
2018-07-04 00:00:00.000 2
2018-07-04 02:00:00.000 1
2018-07-04 06:00:00.000 2
2018-07-04 08:00:00.000 1
I would simply manually create a timetable and join it.
The timetable would have 24 rows (one for each hour) and could look like this:
hour hourtimeframe twohourtimeframe
13 13:00 - 14:00 13:00 - 15:00
I have data in a table with dates, and want to count the rows by "Week of" (e.g., "Week of 2017-05-01"), where the result has the week's date (starting on Mondays) and the count of matching rows — even if there are no rows for that week. (This will all be in a date range.)
I can partition things into weeks readily enough by grouping on DATEPART(wk, D) (where D is the date column), but I'm struggling with:
How to get the "Week of" date and fill, and
How to have a row for a week where there are no matching rows in the data
Here's grouping by week:
SET DATEFORMAT ymd;
SET DATEFIRST 1; -- Monday is first day of week
DECLARE #startDate DATETIME = '2017-05-01';
DECLARE #endDate DATETIME = '2017-07-01';
SELECT DATEPART(wk, D) AS [Week Number], COUNT(*) AS [Count]
FROM #temp
GROUP BY DATEPART(wk, D)
ORDER BY DATEPART(wk, D);
Which gives me:
+−−−−−−−−−−−−−+−−−−−−−+
| Week Number | Count |
+−−−−−−−−−−−−−+−−−−−−−+
| 19 | 5 |
| 20 | 19 |
| 22 | 8 |
| 23 | 10 |
| 24 | 5 |
| 26 | 4 |
+−−−−−−−−−−−−−+−−−−−−−+
But ideally I want:
+−−−−−−−−−−−−+−−−−−−−+
| Week | Count |
+−−−−−−−−−−−−+−−−−−−−+
| 2017-05-01 | 5 |
| 2017-05-08 | 19 |
| 2017-05-15 | 0 |
| 2017-05-22 | 8 |
| 2017-05-29 | 10 |
| 2017-06-05 | 5 |
| 2017-06-12 | 0 |
| 2017-06-19 | 4 |
| 2017-06-26 | 0 |
+−−−−−−−−−−−−+−−−−−−−+
How can I do that?
Set up information for testing:
SET DATEFIRST 1;
SET DATEFORMAT ymd;
CREATE TABLE #temp (
D DATETIME
);
GO
INSERT INTO #temp (D)
VALUES -- Week of 2017-05-01 (#19)
('2017-05-01'),('2017-05-01'),('2017-05-01'),
('2017-05-06'),('2017-05-06'),
-- Week of 2017-05-08 (#20) - note no data actually on the 8th
('2017-05-10'),
('2017-05-11'),('2017-05-11'),('2017-05-11'),('2017-05-11'),('2017-05-11'),('2017-05-11'),
('2017-05-12'),('2017-05-12'),('2017-05-12'),('2017-05-12'),
('2017-05-13'),('2017-05-13'),('2017-05-13'),('2017-05-13'),('2017-05-13'),('2017-05-13'),('2017-05-13'),
('2017-05-14'),
-- Week of 2017-05-15 (#21)
-- (note we have no data for this week)
-- Week of 2017-05-22 (#22)
('2017-05-22'),('2017-05-22'),('2017-05-22'),
('2017-05-23'),('2017-05-23'),('2017-05-23'),('2017-05-23'),('2017-05-23'),
-- Week of 2017-05-29 (#23)
('2017-05-29'),('2017-05-29'),('2017-05-29'),
('2017-06-02'),('2017-06-02'),
('2017-06-03'),
('2017-06-04'),('2017-06-04'),('2017-06-04'),('2017-06-04'),
-- Week of 2017-06-05 (#24) - note no data actually on the 5th
('2017-06-08'),('2017-06-08'),('2017-06-08'),
('2017-06-11'),('2017-06-11'),
-- Week of 2017-06-12 (#25)
-- (note we have no data for this week)
-- Week of 2017-06-19 (#26)
('2017-06-19'),('2017-06-19'),('2017-06-19'),
('2017-06-20');
GO
To do this, you have to generate a table or CTE with the Monday dates and their week numbers (as shown in this answer, slightly modified for what we need to do below), then LEFT JOIN or OUTER APPLY that with your data grouped by week, using the week numbers:
SET DATEFORMAT ymd;
SET DATEFIRST 1;
DECLARE #startDate DATETIME = '2017-05-01';
DECLARE #endDate DATETIME = '2017-07-01';
;WITH Mondays AS (
SELECT #startDate AS D, DATEPART(WK, #startDate) AS W
UNION ALL
SELECT DATEADD(DAY, 7, D), DATEPART(WK, DATEADD(DAY, 7, D))
FROM Mondays m
WHERE DATEADD(DAY, 7, D) < #endDate
)
SELECT LEFT(CONVERT(NVARCHAR(MAX), Mondays.D, 120), 10) AS [Week Of], d.Count
FROM Mondays
OUTER APPLY (
SELECT COUNT(*) AS [Count]
FROM #temp
WHERE DATEPART(WK, D) = W
AND D >= #startDate
AND D < #endDate
) d
ORDER BY Mondays.D;
Two notes on that:
I'm assuming we can ensure that #startDate is a Monday, which is easily done outside the query or could be done with a simple loop in T-SQL if needed (backing up until WEEKPART(WEEKDAY, #startDate) is 1). (Or worst case we could generate all the dates and then filter them with WEEKPART(WEEKDAY, ...).)
I'm assuming the date range is always a year or less; otherwise, we'd have duplicated week numbers. If the date range could be longer than a year, combine the week number with the year everywhere we're just using a week number above (e.g., DATEPART(YEAR, D) * 100 + DATEPART(wk, D)).
You can use this.
SET DATEFORMAT ymd;
SET DATEFIRST 1; -- Monday is first day of week
DECLARE #startDate DATETIME = '2017-05-01';
DECLARE #endDate DATETIME = '2017-07-01';
;WITH OrgResult AS ( -- Grouping result with missing week. Answer of the first question
SELECT
DATEADD(DAY, 1 - DATEPART (WEEKDAY, D), D) [Week] -- Fist Day Of the Week
, COUNT(*) [Count]
FROM #temp
WHERE D BETWEEN #startDate AND #endDate
GROUP BY
DATEADD(DAY, 1 - DATEPART (WEEKDAY, D), D)
)
, Result AS -- Adds only missing weeks. Answer of the second question
(
SELECT * FROM OrgResult
UNION ALL
SELECT DATEADD( DAY, 7, R.[Week] ), 0 [Count]
FROM Result R
WHERE NOT EXISTS( SELECT * FROM OrgResult O WHERE [Week] = DATEADD( DAY, 7, R.[Week] ) )
AND DATEADD( DAY, 7, R.[Week] ) <= #endDate
)
SELECT * FROM Result
ORDER BY [Week]
Result:
Week Count
----------- -----------
2017-05-01 5
2017-05-08 19
2017-05-15 0
2017-05-22 8
2017-05-29 10
2017-06-05 5
2017-06-12 0
2017-06-19 4
2017-06-26 0
Here's another approach. I included this as it will generate less reads than the Recursive CTE Solution and will be a lot fast
WITH E(N) AS (SELECT 1 FROM (VALUES (1),(1),(1),(1),(1),(1),(1),(1),(1),(1))x(x)),
iTally(N) AS
(
SELECT TOP (((DATEDIFF(day,#startdate, #endDate))/7)+1)
(ROW_NUMBER() OVER (ORDER BY (SELECT 1))-1)
FROM E a, E b, E c
)
SELECT WeekOf = DATEADD(WEEK,N,#startDate), [count] = COUNT(t.D)
FROM iTally i
LEFT JOIN #temp t ON t.D >= DATEADD(WEEK,N,#startDate) AND t.D < DATEADD(WEEK,N+1,#startDate)
GROUP BY DATEADD(WEEK,N,#startDate)
ORDER BY DATEADD(WEEK,N,#startDate); -- not required
Results:
WeekOf count
---------- -----------
2017-05-01 5
2017-05-08 19
2017-05-15 0
2017-05-22 8
2017-05-29 10
2017-06-05 5
2017-06-12 0
2017-06-19 4
2017-06-26 0
Consider the below data:
ID Reference Manager LeaseFirstStart LeaseStop
1 KLEIN John 2008-04-02 00:00:00.000 2010-04-01 00:00:00.000
2 HAWKER John 2008-12-18 00:00:00.000 2010-09-17 00:00:00.000
3 SLEEP Bob 2008-01-23 00:00:00.000 2009-01-22 00:00:00.000
4 CODD Bob 2009-08-03 00:00:00.000 2010-08-02 00:00:00.000
5 ALLEN Bob 2008-01-30 00:00:00.000 2009-07-31 00:00:00.000
The earliest month is Jan 2008 and the latest month is Sep 2010.
How can I count the number of leases that were current per month? The output should look like this:
Month Number of Leases
2008-01 2
2008-02 2
2008-03 2
2008-04 3
2008-05 3
2008-06 3
2008-07 3
2008-08 4
… …
Ultimately, I want to use the answer to the question to create the dataset below for use in excel by the user so they can see who had how many leases during the data period.
Month Manager Number of Leases
2008-01 Bob 2
2008-01 John 0
2008-02 Bob 2
2008-02 John 0
2008-03 Bob 2
2008-03 John 0
2008-04 Bob 2
2008-04 John 1
2008-05 Bob 2
2008-05 John 1
2008-06 Bob 2
2008-06 John 1
2008-07 Bob 2
2008-07 John 1
2008-08 Bob 3
2008-08 John 1
… … …
I know I've done it before, but it was a long time ago and I remember it being messy. Thanks in advance!
select sum (no) as no,datet from ( SELECT COUNT (*) as no ,(convert(varchar,datepart (yyyy,[ Start] )) + '-' + convert(varchar, MONTH([ Start] ))) as datet
FROM <tbl>
GROUP BY (convert(varchar,datepart (yyyy,[ Start] )) + '-' + convert(varchar, MONTH([ Start] )))
union SELECT COUNT (*) as no ,(convert(varchar,datepart (yyyy,[ End] )) + '-' + convert(varchar, MONTH([ End] ))) as datet
FROM <tbl>
GROUP BY (convert(varchar,datepart (yyyy,[ End] )) + '-' + convert(varchar, MONTH([ End] )) ) ) t
This is very logical question, finally I created the sql which gives the desired result.. I verified every date and month count and its all ok.
Declare #t table (ID int, Reference varchar(50), Manager varchar(50),LeaseFirstStart datetime,LeaseStop datetime)
insert into #t
values
(1,'KLEIN','John','2008-04-02 00:00:00.000','2010-04-01 00:00:00.000'),
(2,'HAWKER','John','2008-12-18 00:00:00.000','2010-09-17 00:00:00.000'),
(3,'SLEEP','Bob','2008-01-23 00:00:00.000','2009-01-22 00:00:00.000'),
(4,'CODD','Bob','2009-08-03 00:00:00.000','2010-08-02 00:00:00.000'),
(5,'ALLEN','Bob','2008-02-28 00:00:00.000','2009-07-31 00:00:00.000')
declare #lowerdate datetime , #currentdt datetime
select #lowerdate = min(leasefirststart), #currentdt= max(leasestop) from #t
;with cte as
(
select firstday,DATEADD(d, -1, DATEADD(m, DATEDIFF(m, 0, FirstDay) + 1, 0)) Lastday, mng from
( select dateadd(m,datediff(m,0,#lowerdate)+v.number,0) as FirstDay
From master..spt_values v
Where v.type='P' and v.number between 0 and datediff(m, #lowerdate, #currentdt)
) as a
, (select distinct manager mng from #t ) as b
)
select (convert(varchar,datepart (yyyy,FirstDay )) + '-' + convert(varchar, MONTH(FirstDay ))) MonthAndYear ,mng as mng , count( manager ) cnt
from cte
left join #t on
(
firstday between LeaseFirstStart and LeaseStop
or
Lastday between LeaseFirstStart and LeaseStop
) and cte.mng = Manager
group by firstday, mng
order by FirstDay
The time is: (m/d/yyyy) => 2009/01/04
Using this command using datepart(wk,'20090104') I can get the week number (for any given date).
So :
SELECT datepart(wk,'20090101') //1
SELECT datepart(wk,'20090102') //1
SELECT datepart(wk,'20090103') //1
SELECT datepart(wk,'20090104') //2
So far so good.
The problem :
Those 3 first dates are not part of a full week, so I can't put them in a fixed 52-week chart.
Our company needs to see information about each whole week in the 52 weeks of a year. (Each year has 52 whole weeks).
So 20090101 doesn't belong to the first week of 2009 !
It belongs to the previous year (which is irrelevant to my question)
So I need a UDF (I've been searching a lot, and ISOWEEK is not answering my needs) which by a given datetime, will give me the Week Number (week = whole week, so partial weeks aren't considered).
Example :
calcweekNum ('20090101') //52 ...from the last year
calcweekNum ('20090102') //52 ...from the last year
calcweekNum ('20090103') //52 ...from the last year
calcweekNum ('20090104') //1
..
..
calcweekNum ('20090110') //1
calcweekNum ('20090111') //2
calcweekNum ('20090112') //2
...
Here's a different approach. All you need to supply is the year:
DECLARE #year INT = 2009;
DECLARE #start SMALLDATETIME;
SET #start = DATEADD(YEAR, #year-1900, 0);
;WITH n AS
(
SELECT TOP (366) -- in case of leap year
d = DATEADD(DAY, ROW_NUMBER() OVER (ORDER BY name)-1, #start)
FROM sys.all_objects
),
x AS
(
SELECT md = MIN(d) FROM n
WHERE DATEPART(WEEKDAY, d) = 1 -- assuming DATEFIRST is Sunday
),
y(d,wk) AS
(
SELECT n.d, ((DATEPART(DAYOFYEAR, n.d) - DATEDIFF(DAY, #start, x.md)-1)/7) + 1
FROM n CROSS JOIN x
WHERE n.d >= x.md
AND n.d < DATEADD(YEAR, 1, #start)
)
SELECT [date] = d, [week] = wk
FROM y WHERE wk < 53
ORDER BY [date];
Results:
date week
---------- ----
2009-01-04 1
2009-01-05 1
2009-01-06 1
2009-01-07 1
2009-01-08 1
2009-01-09 1
2009-01-10 1
2009-01-11 2
2009-01-12 2
...
2009-12-25 51
2009-12-26 51
2009-12-27 52
2009-12-28 52
2009-12-29 52
2009-12-30 52
2009-12-31 52
Note that week 52 won't necessarily be a full week, and that in some cases (e.g. 2012), the last day or two of the year might fall in week 53, so they're excluded.
An alternative approach is to repeat the MIN expression twice:
DECLARE #year INT = 2009;
DECLARE #start SMALLDATETIME;
SET #start = DATEADD(YEAR, #year-1900, 0);
;WITH n AS
(
SELECT TOP (366) -- in case of leap year
d = DATEADD(DAY, ROW_NUMBER() OVER (ORDER BY name)-1, #start)
FROM sys.all_objects
),
y(d,wk) AS
(
SELECT n.d, ((DATEPART(DAYOFYEAR, n.d) - DATEDIFF(DAY, #start, (SELECT MIN(d)
FROM n WHERE DATEPART(WEEKDAY, d) = 1))-1)/7) + 1
FROM n
WHERE n.d >= (SELECT md = MIN(d) FROM n WHERE DATEPART(WEEKDAY, d) = 1)
AND n.d < DATEADD(YEAR, 1, #start)
)
SELECT [date] = d, [week] = wk
FROM y WHERE wk < 53
ORDER BY d;
Here's a function for you to calculate it on the fly:
CREATE FUNCTION dbo.WholeWeekFromDate (
#Date datetime
)
RETURNS tinyint
AS BEGIN
RETURN (
SELECT DateDiff(Day, DateAdd(Year, DateDiff(Year, 0, CalcDate), 0), CalcDate) / 7 + 1
FROM (SELECT DateAdd(Day, (DateDiff(Day, 0, #Date) + 1) / 7 * 7, 0)) X (CalcDate)
);
END;
I don't recommend you use it, as it may perform badly due to being called once for every row. If you absolutely must have a function to use in real queries, then convert it to an inline function returning a single column and row, and use it as so:
SELECT
OtherColumns,
(SELECT WeekNumber FROM dbo.WholeWeekFromDate(DateColumn)) WeekNumber
FROM
YourTable;
This will allow it to be "inlined" in the execution plan and perform significantly better.
But even better, as others have suggested, is to use a BusinessDate table. Here's a head start on creating one for you:
CREATE TABLE dbo.BusinessDate (
BusinessDate date NOT NULL CONSTRAINT PK_BusinessDate PRIMARY KEY CLUSTERED,
WholeWeekYear smallint NOT NULL
CONSTRAINT CK_BusinessDate_WholeWeekYear_Valid
CHECK (WholeWeekYear BETWEEN 1900 AND 9999),
WholeWeekNumber tinyint NOT NULL
CONSTRAINT CK_BusinessDate_WholeWeekNumber_Valid
CHECK (WholeWeekNumber BETWEEN 1 AND 53),
Holiday bit CONSTRAINT DF_BusinessDate_Holiday DEFAULT (0),
Weekend bit CONSTRAINT DF_BusinessDate_Weekend DEFAULT (0),
BusinessDay AS
(Convert(bit, CASE WHEN Holiday = 0 AND Weekend = 0 THEN 1 ELSE 0 END)) PERSISTED
);
And I'll even populate it from 1900-01-01 through 2617-09-22 (Is that enough for the projected life of your product? And it's only 7.8MB so don't fret over size):
WITH A (N) AS (SELECT 1 UNION ALL SELECT 1 UNION ALL SELECT 1 UNION ALL SELECT 1
UNION ALL SELECT 1 UNION ALL SELECT 1 UNION ALL SELECT 1 UNION ALL SELECT 1),
B (N) AS (SELECT 1 FROM A F, A A, A L, A C, A O, A N),
C (N) AS (SELECT Row_Number() OVER (ORDER BY (SELECT 1)) FROM B),
Dates AS (
SELECT
N,
DateAdd(Day, N, '18991231') Dte,
DateAdd(Day, N / 7 * 7, '19000101') CalcDate
FROM C
)
INSERT dbo.BusinessDate
SELECT
Dte,
Year(CalcDate),
DateDiff(Day, DateAdd(Year, DateDiff(Year, 0, CalcDate), 0), CalcDate) / 7 + 1,
0,
(N + 6) % 7 / 5 -- calculate weekends
FROM Dates; -- 3-7 seconds or so on my VM server
Then join to the table on the date, and use the WholeWeekNumber column for your output. You might also consider adding a WeekNumberYear because it's going to be a tad difficult to figure out that the 52 of 2009-01-01 really belongs to 2008 without this... a strange data point in there for sure if you don't (laugh).
Example table contents:
BusinessDate WholeWeekYear WholeWeekNumber Holiday Weekend BusinessDay
------------ ------------- --------------- ------- ------- -----------
1/1/2009 2008 52 0 0 1
1/2/2009 2008 52 0 0 1
1/3/2009 2008 52 0 1 0
1/4/2009 2009 1 0 1 0
1/5/2009 2009 1 0 0 1
1/6/2009 2009 1 0 0 1
1/7/2009 2009 1 0 0 1
1/8/2009 2009 1 0 0 1
1/9/2009 2009 1 0 0 1
1/10/2009 2009 1 0 1 0
1/11/2009 2009 2 0 1 0
If you really don't want to use this as a general business date calculation table, you can drop the last 3 columns, otherwise, update the Holiday column to 1 for company holidays.
Note: if you actually make the above table, and your access to it most often uses JOIN or WHERE conditions on a different column than BusinessDate, then make the primary key nonclustered and add a clustered index starting with the alternate column.
Some of the above scripts require SQL 2005 or higher.
It would be relatively easy to setup a custom calendar table with one row for each date of the year in it, and then have other fields that will allow you to rollup however you want. I do this when I have clients using varying calendars, i.e. fiscal years, and it makes the query logic very simple.
Then you just join date-to-date and get the week-number that you want.
date | reporting year | reporting week
-----------|----------------|---------------
2009-01-01 | 2008 | 52
2009-01-02 | 2008 | 52
2009-01-03 | 2008 | 52
2009-01-04 | 2009 | 01
2009-01-05 | 2009 | 01
etc.
and then to use it ( for example to get total sales rollup by your custom weeks, didn't validated my sql):
select reporting_year, reporting_month, sum(sales)
from sales
inner join custom_date_table cdt on cdt.sysdate = sales.sysdate
group by reporting_year, reporting_month
where report_year=2009
DECLARE #StartDate DATE;
SET #StartDate = '20120101';
WITH Calendar AS (
SELECT #StartDate AS DateValue
,DATEPART(DW, #StartDate) AS DayOfWeek
,CASE WHEN DATEPART(DW, #StartDate) = 1 THEN 1 ELSE 0 END AS WeekNumber
UNION ALL
SELECT DATEADD(d, 1, DateValue)
,DATEPART(DW, DATEADD(d, 1, DateValue)) AS DayOfWeek
,CASE WHEN DayOfWeek = 7 THEN WeekNumber + 1 ELSE WeekNumber END
FROM Calendar
WHERE DATEPART(YEAR, DateValue) = DATEPART(YEAR, #StartDate)
)
SELECT DateValue, WeekNumber
FROM Calendar
WHERE WeekNumber BETWEEN 1 AND 52
AND DATEPART(YEAR, DateValue) = DATEPART(YEAR, #StartDate)
OPTION (MAXRECURSION 0);
Don't use a UDF, use a calendar table instead, then you can define week numbers exactly as your company requires and simply query them from the table, which will be much easier and possibly much faster than using a UDF.
A calendar table has numerous uses in SQL (search this site or Google) so you should probably create it anyway.
There is no good answer for this.
A year is NOT 52 weeks long.
It is 52 weeks and one day in normal years, and 52 weeks and two days in leap years.
I have a table containing a number of timestamps per day, they represents start and stop events.
ID TimeStamp
----------------------
1 2008-01-01 07:00:00
1 2008-01-01 08:15:00
1 2008-01-01 10:00:00
1 2008-01-01 11:00:00
1 2008-01-02 10:30:00
1 2008-01-02 12:00:00
I would like to calcuate the total running time per day, like this:
ID Date RunningTime
-------------------------
1 2008-01-01 02:15:00
1 2008-01-02 01:30:00
Do anyone have a nice T-SQL solution for my problem?
WITH q AS
(
SELECT *,
CONVERT(DATETIME, CONVERT(VARCHAR(8), TimeStamp, 112), 112) AS dte,
ROW_NUMBER() OVER (PARTITION BY id, CONVERT(DATETIME, CONVERT(VARCHAR(8), TimeStamp, 112), 112) ORDER BY TimeStamp) AS rn
FROM mytable
)
SELECT qb.id, qb.dte, SUM(DATEDIFF(second, qb.TimeStamp, qe.TimeStamp))
FROM q qb
JOIN q qe
ON qe.id = qb.id
AND qe.dte = qb.dte
AND qe.rn = qb.rn + 1
WHERE qb.rn % 2 = 1
GROUP BY
qb.id, qb.dte
This assumes that every record open on a certain day should also be closed on the same day.