I want to add 30 consecutive days of data in my Date Dimension table using DATEDIFF() but I am getting blank result. Can you please help me correct the code below to get the desired result?
CREATE TABLE dbo.dateDimension (
DateKey INT NOT NULL
,DateValue DATE NOT NULL
,CYear SMALLINT NOT NULL
,CMonth TINYINT NOT NULL
,CONSTRAINT PK_DimDate PRIMARY KEY ( DateKey )
);
GO
CREATE PROC dbo.dateTest
#StartDate DATETIME
AS
WHILE (DATEDIFF(day, #StartDate, GETDATE()) <=30)
BEGIN
INSERT into dbo.dateDimension
SELECT CAST( YEAR(#StartDate) * 10000 + MONTH(#StartDate) * 100 + DAY(#StartDate) AS INT)
,#StartDate
,YEAR(#StartDate)
,MONTH(#StartDate)
SET #StartDate = DATEADD(d,1,#StartDate)
END;
GO
EXECUTE dbo.dateTest '2010-01-01'
SELECT * FROM dbo.dateDimension
The issue is that this logic:
DATEDIFF(day, #StartDate, GETDATE())
gives 3739 days with your current start date, so its never less than 30. Personally I would simply count it as follows:
DECLARE #StartDate DATETIME = '2010-01-01', #Count INT = 0;
WHILE #Count <= 30 BEGIN
INSERT into dbo.dateDimension
SELECT CAST( YEAR(#StartDate) * 10000 + MONTH(#StartDate) * 100 + DAY(#StartDate) AS INT)
, #StartDate
, YEAR(#StartDate)
, MONTH(#StartDate);
SET #StartDate = DATEADD(d,1,#StartDate);
set #Count = #Count + 1;
END;
SELECT *
FROM dbo.dateDimension;
If you are using SQL Server 2016 or above, this solution will not use a while loop, instead it uses a CTE to generate 30 rows numbered I to 30 and then uses the date to convert to yyyymmdd.
DECLARE #NUM_DAYS INT=30;
DECLARE #STARTDATE DATETIME='2020-01-01';
WITH CTE AS(
SELECT ROW_NUMBER() OVER(ORDER BY (SELECT NULL)) AS SRL
FROM STRING_SPLIT(REPLICATE(',',#num_days-1), ',') AS A
)
INSERT INTO dbo.dateDimension
SELECT
CONVERT(INT, CONVERT(CHAR(8), DATEADD(DAY, SRL-1, #STARTDATE), 112))
, #STARTDATE
, YEAR(#STARTDATE)
, MONTH(#STARTDATE)
FROM CTE
SELECT * FROM dbo.dateDimension
I have customer_since stored as YYYYWW i.e., 201852.
The below script is what I have been using to work out the difference, however when I have this situation of 201901 - 201852, it gives 48 instead of 1.
is there mod function or something that can be incorporated here to resolve the issue?
CAST(CONCAT (DATEPART(YEAR, DATEADD(MONTH, + 9, GETDATE()))
, RIGHT('0' + RTRIM(DATEPART(WEEK, DATEADD(MONTH, - 3, GETDATE()))+1),2)
) AS INT) - a.customer_since AS Customer_since
You can use the following template to get the first (or the last) date of a given year and week. Then having valid date times, simply use DATEDIFF as it has built in functionality to get difference in weeks:
DECLARE #WeekNo int= 52
DECLARE #Year int=2018
SELECT DATEADD(wk,#WeekNo-1,DATEADD(yy,#Year-1900,0)) AS WeekStart,
DATEADD(wk,#WeekNo,DATEADD(yy,#Year-1900,0))-1 AS WeekEnd
GO
DECLARE #WeekNo int= 1
DECLARE #Year int=2019
SELECT DATEADD(wk,#WeekNo-1,DATEADD(yy,#Year-1900,0)) AS WeekStart,
DATEADD(wk,#WeekNo,DATEADD(yy,#Year-1900,0))-1 AS WeekEnd
GO
SELECT DATEDIFF(WEEK, '2018-12-24 00:00:00.000', '2019-01-01 00:00:00.000')
In one statement:
DECLARE #Input01 VARCHAR(6) = '201852'
,#Input02 VARCHAR(6) = '201901'
SELECT DATEDIFF(WEEK, DATEADD(WEEK, RIGHT(#Input01, 2)-1,DATEADD(YEAR,LEFT(#Input01, 4)-1900,0)), DATEADD(WEEK,RIGHT(#Input02, 2)-1,DATEADD(YEAR,LEFT(#Input02, 4)-1900,0)));
You can create a function which will accept the date string('201852') and will return a date like `Create Function Todate(#dt varchar(100))
returns Date
Begin
Declare #wk int=(Select RIGHT(#dt,2))
return (Select Dateadd(WK, #wk, cast(LEFT(#dt,4) as date)))
End`
Now all you have to do is pass the date string in the function and find a DATEDIFF with start and End date
I'm working on a query that deals with a frequency value (i.e. Mondays, Tuesdays, etc. - Think assignments).
So in my query I currently have a result of
jobId:1, personId:100, frequencyVal: 'Mondays'
jobId:2, personId:101, frequencyVal: 'Saturdays'
What I need is the next the 4 future(or current) dates for the frequencyVal.
So if today is 1/3/2015
I would need my result set to be
jobId:1, personId:100, frequencyVal: 'Mondays', futureDates: '1/5,1/12,1/19,1/26'
jobId:2, personId:102, frequencyVal: 'Saturdays', futureDates: '1/3,1/10,1/17,1/24'
I was looking at the following post:
How to find the Nearest (day of the week) for a given date
But that sets it for a specific date. And I'm looking at this being a web application and I want the dates for the current date. So if I try to run this query next Tuesday the future dates for jobId:1 would remove the 1/5 and add the 2/2.
Is there a way to pass in a weekday value to get the next nearest date?
I prefer a calendar table for this kind of query. Actually, I prefer a calendar table over date functions for most queries. Here's a minimal one. The one I use in production has more columns and more rows. (100 years of data is only 37k rows.)
create table calendar (
cal_date date not null primary key,
day_of_week varchar(15)
);
insert into calendar (cal_date) values
('2015-01-01'), ('2015-01-02'), ('2015-01-03'), ('2015-01-04'),
('2015-01-05'), ('2015-01-06'), ('2015-01-07'), ('2015-01-08'),
('2015-01-09'), ('2015-01-10'), ('2015-01-11'), ('2015-01-12'),
('2015-01-13'), ('2015-01-14'), ('2015-01-15'), ('2015-01-16'),
('2015-01-17'), ('2015-01-18'), ('2015-01-19'), ('2015-01-20'),
('2015-01-21'), ('2015-01-22'), ('2015-01-23'), ('2015-01-24'),
('2015-01-25'), ('2015-01-26'), ('2015-01-27'), ('2015-01-28'),
('2015-01-29'), ('2015-01-30'), ('2015-01-31'),
('2015-02-01'), ('2015-02-02'), ('2015-02-03'), ('2015-02-04'),
('2015-02-05'), ('2015-02-06'), ('2015-02-07'), ('2015-02-08'),
('2015-02-09'), ('2015-02-10'), ('2015-02-11'), ('2015-02-12'),
('2015-02-13'), ('2015-02-14'), ('2015-02-15'), ('2015-02-16'),
('2015-02-17'), ('2015-02-18'), ('2015-02-19'), ('2015-02-20'),
('2015-02-21'), ('2015-02-22'), ('2015-02-23'), ('2015-02-24'),
('2015-02-25'), ('2015-02-26'), ('2015-02-27'), ('2015-02-28')
;
update calendar
set day_of_week = datename(weekday, cal_date);
alter table calendar
alter column day_of_week varchar(15) not null;
alter table calendar
add constraint cal_date_matches_dow
check (datename(weekday, cal_date) = day_of_week);
create index day_of_week_ix on calendar (day_of_week);
Set the privileges so that
everyone can select, but
almost nobody can insert new rows, and
even fewer people can delete rows.
(Or write a constraint that can guarantee there are no gaps. I think you can do that in SQL Server.)
You can select the next four Mondays after today with a very simple SQL statement. (The current date is 2015-01-05, which is a Monday.)
select top 4 cal_date
from calendar
where cal_date > convert(date, getdate())
and day_of_week = 'Monday'
order by cal_date;
CAL_DATE
--
2015-01-12
2015-01-19
2015-01-26
2015-02-02
For me, this is a huge advantage. No procedural code. Simple SQL that is obviously right. Big win.
Your sample table
create table #t
(
jobId int,
personId int,
frequencyVal varchar(10)
);
insert into #t values (1,100,'Mondays'),(2,101,'Saturdays');
QUERY 1 : Select nearest 4 week of days in current month for particular week day
-- Gets first day of month
DECLARE #FIRSTDAY DATE=DATEADD(month, DATEDIFF(month, 0, GETDATE()), 0)
;WITH CTE as
(
-- Will find all dates in current month
SELECT CAST(#FIRSTDAY AS DATE) as DATES
UNION ALL
SELECT DATEADD(DAY,1,DATES)
FROM CTE
WHERE DATES < DATEADD(MONTH,1,#FIRSTDAY)
)
,CTE2 AS
(
-- Join the #t table with CTE on the datename+'s'
SELECT jobId,personId,frequencyVal,DATES,
-- Get week difference for each weekday
DATEDIFF(WEEK,DATES,GETDATE()) WEEKDIFF,
-- Count the number of weekdays in a month
COUNT(DATES) OVER(PARTITION BY DATENAME(WEEKDAY,CTE.DATES)) WEEKCOUNT
FROM CTE
JOIN #t ON DATENAME(WEEKDAY,CTE.DATES)+'s' = #t.frequencyVal
WHERE MONTH(DATES)= MONTH(GETDATE())
)
-- Converts to CSV and make sure that only nearest 4 week of days are generated for month
SELECT DISTINCT C2.jobId,C2.personId,frequencyVal,
SUBSTRING(
(SELECT ', ' + CAST(DATEPART(MONTH,DATES) AS VARCHAR(2)) + '/' +
CAST(DATEPART(DAY,DATES) AS VARCHAR(2))
FROM CTE2
WHERE C2.jobId=jobId AND C2.personId=personId AND C2.frequencyVal=frequencyVal AND
((WEEKDIFF<3 AND WEEKDIFF>-3 AND WEEKCOUNT = 5) OR WEEKCOUNT <= 4)
ORDER BY CTE2.DATES
FOR XML PATH('')),2,200000) futureDates
FROM CTE2 C2
SQL FIDDLE
For example, in Query2 the nearest date(here we take example as Saturday) of
2015-Jan-10 will be 01/03,01/10,01/17,01/24
2015-Jan-24 will be 01/10,01/17,01/24,01/31
QUERY 2 : Select next 4 week's dates for particular week day irrelevant of month
;WITH CTE as
(
-- Will find the next 4 week details
SELECT CAST(GETDATE() AS DATE) as DATES
UNION ALL
SELECT DATEADD(DAY,1,DATES)
FROM CTE
WHERE DATES < DATEADD(DAY,28,GETDATE())
)
,CTE2 AS
(
-- Join the #t table with CTE on the datename+'s'
SELECT jobId,personId,frequencyVal, DATES,
ROW_NUMBER() OVER(PARTITION BY DATENAME(WEEKDAY,CTE.DATES) ORDER BY CTE.DATES) DATECNT
FROM CTE
JOIN #t ON DATENAME(WEEKDAY,CTE.DATES)+'s' = #t.frequencyVal
)
-- Converts to CSV and make sure that only 4 days are generated for month
SELECT DISTINCT C2.jobId,C2.personId,frequencyVal,
SUBSTRING(
(SELECT ', ' + CAST(DATEPART(MONTH,DATES) AS VARCHAR(2)) + '/' +
CAST(DATEPART(DAY,DATES) AS VARCHAR(2))
FROM CTE2
WHERE C2.jobId=jobId AND C2.personId=personId AND C2.frequencyVal=frequencyVal
AND DATECNT < 5
ORDER BY CTE2.DATES
FOR XML PATH('')),2,200000) futureDates
FROM CTE2 C2
SQL FIDDLE
The following would be the output if the GETDATE() (if its Saturday) is
2015-01-05 - 1/10, 1/17, 1/24, 1/31
2015-01-24 - 1/24, 1/31, 2/7, 2/14
There's no built-in function to do it. But you can try this, you may place it inside a Scalar-Valued Function:
DECLARE #WeekDay VARCHAR(10) = 'Monday';
DECLARE #WeekDayInt INT;
SELECT #WeekDayInt = CASE #WeekDay
WHEN 'SUNDAY' THEN 1
WHEN 'MONDAY' THEN 2
WHEN 'TUESDAY' THEN 3
WHEN 'WEDNESDAY' THEN 4
WHEN 'THURSDAY' THEN 5
WHEN 'FRIDAY' THEN 6
WHEN 'SATURDAY' THEN 7 END
SELECT CONVERT(DATE, DATEADD(DAY, (DATEPART(WEEKDAY, GETDATE()) + #WeekDayInt) % 7, GETDATE())) AS NearestDate
UPDATE:
Looks like radar was right, here's the solution:
DECLARE #WeekDay VARCHAR(10) = 'Monday';
DECLARE #WeekDayInt INT;
DECLARE #Date DATETIME = GETDATE();
SELECT #WeekDayInt = CASE #WeekDay
WHEN 'SUNDAY' THEN 1
WHEN 'MONDAY' THEN 2
WHEN 'TUESDAY' THEN 3
WHEN 'WEDNESDAY' THEN 4
WHEN 'THURSDAY' THEN 5
WHEN 'FRIDAY' THEN 6
WHEN 'SATURDAY' THEN 7 END
DECLARE #Diff INT = DATEPART(WEEKDAY, #Date) - #WeekDayInt;
SELECT CONVERT(DATE, DATEADD(DAY, CASE WHEN #Diff >= 0 THEN 7 - #Diff ELSE ABS(#Diff) END, #Date)) AS NearestDate
Try this - based on king.code's answer to get the nearest date.
create table #t
(
jobId int,
personId int,
frequencyVal varchar(10)
);
insert into #t values (1,100,'Mondays'),(2,101,'Saturdays');
WITH cte(n) AS
(
SELECT 0
UNION ALL
SELECT n+1 FROM cte WHERE n < 3
)
select #t.jobId, #t.personId, #t.frequencyVal, STUFF(a.d, 1, 1, '') AS FutureDates
from #t
cross apply (SELECT CASE #t.frequencyVal
WHEN 'SUNDAYS' THEN 1
WHEN 'MONDAYS' THEN 2
WHEN 'TUESDAYS' THEN 3
WHEN 'WEDNESDAYS' THEN 4
WHEN 'THURSDAYS' THEN 5
WHEN 'FRIDAYS' THEN 6
WHEN 'SATURDAYS' THEN 7
END)tranlationWeekdays(n)
cross apply (select ',' + CONVERT(varchar(10), CONVERT(date,dateadd(WEEK, cte.n,CONVERT(DATE, DATEADD(DAY, (DATEPART(WEEKDAY, GETDATE()) + tranlationWeekdays.n) % 7, GETDATE()))))) from cte FOR XML PATH('')) a(d);
drop table #t;
Try this,
DECLARE #YEAR INT=2015
DECLARE #MONTH INT=1
DECLARE #DAY INT=1
DECLARE #DATE DATE = (SELECT DateFromParts(#Year, #Month, #Day))
DECLARE #TOTAL_DAYS INT =(SELECT DatePart(DY, #DATE));
WITH CTE1
AS (SELECT T_DAY=(SELECT DateName(DW, #DATE)),
#DATE AS T_DATE,
#DAY AS T_DDAY
UNION ALL
SELECT T_DAY=(SELECT DateName(DW, DateAdd(DAY, T_DDAY + 1, #DATE))),
DateAdd(DAY, T_DDAY + 1, #DATE) AS T_DATE,
T_DDAY + 1
FROM CTE1
WHERE T_DDAY + 1 <= 364)
SELECT DISTINCT T_DAY,
Stuff((SELECT ',' + CONVERT(VARCHAR(30), T_DATE)
FROM CTE1 A
WHERE A.T_DAY=CTE1.T_DAY AND A.T_DATE > GetDate() AND A.T_DATE<(DATEADD(WEEK,4,GETDATE()))
FOR XML PATH('')), 1, 1, '') AS FUTURE
FROM CTE1
ORDER BY T_DAY
OPTION (MAXRECURSION 365)
This is a simpler way I think, and I think it fits your requirements.
Note that I have changed your frequency_val column to an integer that represents the day of the week from SQL servers perspective and added a calculated column to illustrate how you can easily derive the day name from that.
declare #t table
(
jobId int,
personId int,
--frequencyVal varchar(10)
frequency_val int,
frequency_day as datename(weekday,frequency_val -1) + 's'
);
declare #num_occurances int = 4
declare #from_date date = dateadd(dd,3,getdate()) -- this will allow you to play with the date simply by changing the increment value
insert into #t
values
(1,100,1),--'Mondays'),
(2,101,6),--'Saturdays');
(3,101,7),--'Saturdays');
(4,100,2)--'Mondays'),
--select * from #t
;with r_cte (days_ahead, occurance_date)
as (select 0, convert(date,#from_date,121)
union all
select r_cte.days_ahead +1, convert(date,dateadd(DD, r_cte.days_ahead+1, #from_date),121)
from r_cte
where r_cte.days_ahead < 7 * #num_occurances
)
select t.*, r_cte.occurance_date
from
#t t
inner join r_cte
on DATEPART(WEEKDAY, dateadd(dd,##DATEFIRST - 1 ,r_cte.occurance_date)) = t.frequency_val
Having seen the use of DATENAME in some of the answers already given, I'd like to point out that return values of DATENAME might vary depending on your current language setting, but you can save the current language setting and ensure usage of us_english so you can be confident to use English weekday names.
Now here is my slightly different approach to get the 4 next dates that fall on a certain (known) weekday, using a user defined table valued function that allows to create a number sequence table (yes this is a pretty dull function, you have to pass MaxValue greater MinValue, but that could be easily enhanced, if needed, but hey, it does the job). Using that function span a table over 28 values (next 28 days should indeed include the next 4 relevant weekdays ;)), apply DATEADD on GETDATE and reduce the result set with WHERE to only those values that have the right weekday:
CREATE FUNCTION GetIntSequence(#MinValue INT, #MaxValue INT)
RETURNS #retSequence TABLE
(
IntValue INT NOT NULL
)
BEGIN
DECLARE #i INT = (SELECT #MinValue)
WHILE #i <= #MaxValue
BEGIN
INSERT INTO #retSequence (IntValue) SELECT #i
SELECT #i = #i + 1
END
RETURN
END
GO
DECLARE #weekDay NVARCHAR(MAX) = 'Monday' --(or Tuesday, wednesday, ...)
--save current language setting
DECLARE #languageBackup NVARCHAR(MAX) = (SELECT ##LANGUAGE)
--ensure us english language setting for reliable weekday names
SET LANGUAGE us_english;
SELECT FourWeeks.SomeDay FROM
(
SELECT
DATEADD(DAY, IntValue, GETDATE()) AS SomeDay
FROM dbo.GetIntSequence(1, 28)
) AS FourWeeks
WHERE DATENAME(WEEKDAY, SomeDay) = #weekDay
--restore old language setting
SET LANGUAGE #languageBackup;
GO
DROP FUNCTION dbo.GetIntSequence
I have a table in which joining dates are give in datetime format.
I have to calculate how many employees joined each financial year resp. ie for eg from
1-04-2002 to 31-03-2003.this should work for each year..from 2003 to 2004,2004 to 2005...n so on.
can anybdy help?
thanxx.
You can map start date to financial year using YEAR(DATEADD(M,-3,JoinDate) I think and you can count records with a CTE, e.g.
with EmployeeStartFinYear(FinYear, EmployeeId)
as
(
select year(dateadd(M,-3,JoinDate)), EmployeeId
from Employees
where JoinDate is not null
)
select FinYear, count(EmployeeId)
from EmployeeStartFinYear
group by FinYear
order by FinYear;
Here's my answer. I think it looks horrid, but i think it works. I'll explain the logic behind it.
I declared the Start and End dates just because it's easier to write than a date.
The #Years variable is the difference in years between the start and end date.
#Counter is used to loop through the number of years stored in #Years. #Diff is always one more than #Counter because each time we go through the loop, we want to increment the date range so it's always 1 year, rather than be counting employees that joined in 1 year, then 2 years etc.
#TempTable stores the info we get each time we go through the query.
All the query does is get the count of employees between the Start Date and a year from that start date and puts it into a temp table. Then it looks through again, and gets the employees that started between Start Date + 1 and Start Date + 2.
Sorry if it's horrible and ugly and doesn't work.
DECLARE #StartDate DATETIME
DECLARE #EndDate DATETIME
DECLARE #Years TINYINT
DECLARE #Counter TINYINT
DECLARE #Diff TINYINT
DECLARE #TempTable TABLE
(
FinancialYear VARCHAR(9)
,Employees TINYINT
)
SET #Count = 0
SET #Diff = 1
SET #Years = DATEDIFF(yyyy, #StartDate, #EndDate)
WHILE #Count < #Years - 1
BEGIN
SELECT
CAST(DATEPART(yyyy, DATEADD(yyyy, #Count, #StartDate) AS VARCHAR(4)) + '-' + CAST(DATEPART(yyyy, DATEADD(yyyy, #Diff, #StartDate)) AS VARCHAR(4) AS FinancialYear
,COUNT(employee_id) AS Employees
INTO #TempTable
FROM
Employees
WHERE
join_date >= #StartDate AND join_date < DATEADD(yyyy, 1, #StartDate)
GROUP BY
CAST(DATEPART(yyyy, DATEADD(yyyy, #Count, #StartDate) AS VARCHAR(4)) + '-' + CAST(DATEPART(yyyy, DATEADD(yyyy, #Diff, #StartDate)) AS VARCHAR(4)
SET #Count = #Count + 1
SET #Diff = #Diff + 1
END
SELECT * FROM #TempTable