Surely there is a way to store a date value prior than 1-1-1900 in a column other than varchar? Or do I really have to break this out, and perform my own datetime library for date differences, sorting, and such?
Yes, I understand how datetime's are actually stored (an integer value from a specific date), but I remember there being another workaround from years ago. It's just slipping me now.
smalldatetime vs. datetime!
http://msdn.microsoft.com/en-us/library/ms187819.aspx
http://msdn.microsoft.com/en-us/library/ms182418.aspx
!
Yes the SQL Server DateTime type can store dates from 1 January 1753.
SQL Server 2008 has the DATE data type which can range from 0001-01-01 through 9999-12-31.
With previous versions, your SOL if you need a date prior to 1753-01-01 (Minimum value for DATETIME).
Related
In my aplication I use a DateEdit control and show the date using this format dd/MM/yyyy so the date is save correctly when I use mssql-server 2014 (spanish) but whe I use mssql-server 2008 (english) the date is stored with this fomat MM/dd/yyyy so what I need to know is if there is a way to store the date with the format dd/MM/yyyy no matter the versión of mssql-server I use or the lenguage I use?
EDIT: to explain better this is the problem, this text 03/07/2017 is stored in the sqlserver 2014 as july 3 of 2017 and in the sqlserver 2008 is stored as march 7 of 2017.
Convert.ToDateTime(String).ToString("dd-MM-yyyy")
Check the defualt language on your MS SQL Server 2008, and if its possible, you can change to Spanish: How to change Date Format after installing SQL Server
Also, you can format your query output with the CONVERT function (#ZLK post the link) using the format 103: CONVERT (DATETIME,[YOURDATE],103)
Following solution has two options. You can use one as per your requirements.
DECLARE #Date AS TABLE(val VARCHAR(30));
INSERT INTO #Date VALUES('12/04/2016');
INSERT INTO #Date VALUES('10/01/2015');
INSERT INTO #Date VALUES('17/03/2017');
--Option 1
SELECT TRY_PARSE(val AS DATE USING 'en-gb') 'Option 1' FROM #Date;
--Option 2
SELECT CONVERT(VARCHAR(10),TRY_PARSE(val AS DATE USING 'en-gb'),103) 'Option 2' FROM #Date;
And the outputs are:
As others said: SQL doe not store dates as we humans do. But thats another matter.
I think what you need is to inform SQL the format your date has.
The best option is to use ISO 8601 format: yyyy-mm-ddThh:mi:ss.mmm.
So, if before saving the date you just format it with this format, SQL (no matter the version or the language) will recognize it as begining with a year, followed by a month and so on. The key of this subject is using first 4 digits.
This way, 2017-01-03 will always be January 03, 2017 and never March 01, 2017. Hope this help you.
Using this you can change the date Format
select FORMAT(GETDATE(), 'dd/MMM/yyyy', 'en-us')
Are there any ways to store the datetime type that goes outside of the type range? I mean if i try to do smth likeINSERT INTO [tbl](..., [IndependenceDate],....) VALUES(...,'1650-01-01',...) i will get an error.(cause datetime type range is: January 1, 1753, through December 31, 9999). I am using SQL Server Management Studio.
You can use Date if you don't care about the time and have a newer version of SQL. Alternatively you can use DateTime2 (but this will use more space)...
I'm pulling the data from SQL database. I have a couple columns with date which need to be converted into Int type, but when I do this the date changes (-2 days). I tried Cast and Convert and it's always the same.
Converting to other type works fine and returns the correct date, but doesn't work for me. I need only the date part from datetime and it needs to be recognised as a date by Excel.
Why is this happening? Any ideas how to get it sorted?
I'm using the following query:
SELECT wotype3, CONVERT(INT,wo_date2 ,103), CAST(duedate AS int) FROM Tasks WHERE
duedate > DATEADD(DAY,1, GETDATE())
AND wo_date2>0
AND wo_date2<DATEADD(WEEK,3,GETDATE())
ORDER BY wotype3
I've had big problems with this, checking my SQL Server's calculation results with "expected results" which a user had created using Excel.
We had discrepancies just because of this 2-day date difference.
Why does it happen ?
Two reasons:
SQL Server uses a zero-based date count from Jan 1 1900, but Excel uses a 1-based date count from Jan 1 1900.
Excel has a bug in it (gasp!) which makes it think that the year 1900 was a leap year. It wasn't. SQL Server correctly refuses to let you have a date value containing "29-Feb-1900".
Combine these two discrepancies, and this is why all dates, from March 1 1900 onwards, are always 2-days out.
Apparently, this Excel bug is a known issue, to keep it in line with Lotus 1-2-3.
The Intentional Date Bug
Microsoft's own explanation
From now on, I think I'll justify bugs in my code with the same excuse.
;-)
For SQL Server 2008 and above, you can use the DATE datatype.
declare #dt datetime = '12/24/2013 10:45 PM' -- some date for example
SELECT #dt as OriginalDateTime, CAST(#dt as DATE) as OnlyDate
For versions prior to SQL Server 2008, you would need to truncate the time part using one or the other functions. Here is one way to do that:
declare #dt datetime = '12/24/2013 10:45 PM' -- some date for example
SELECT #dt as OriginalDateTime, CAST(FLOOR(CAST(#dt AS FLOAT)) as DATETIME) as OnlyDate
if any date time value provided to sql server can i get it's midnight value with some function in sql server.. for example if i provide 2013/07/03 01:34AM , i want to get it to 2013/07/03 12:00 AM.Is there a way to do it?
SQL Server 2008+
SELECT CAST(CAST('2013/07/03 01:34AM' AS date) AS datetime)
For older versions, see this Best approach to remove time part of datetime in SQL Server Never use anything that requires float or int or varchar conversions
This should give you what you need:
SELECT DATEADD(DAY, DATEDIFF(DAY, 0, InputDateField), 0)
Should be slightly quicker than cast:
Most efficient way in SQL Server to get date from date+time?
I'm currently using
Convert(varchar, Getdate(), 101)
to insert only date part of system date into one of my sql server database tables.
my question is: is it the right way to do that or is there any other better method to do it?
I don't understand why you're converting the GETDATE() output (which is DATETIME already) to a VARCHAR and then SQL Server would convert it back to DATETIME upon inserting it again.
Just use:
INSERT INTO dbo.YourTable(SomeDateTimeColumn)
VALUES(GETDATE())
If you're doing that conversion to get rid of the time portion of the DATETIME, you should better:
use the DATE datatype (available in SQL Server 2008 and newer) to store only the DATE (no time)
if you're using SQL Server 2005 or earlier, use this conversion instead - should be much more efficient than two conversions!
INSERT INTO dbo.YourTable(SomeDateTimeColumn)
VALUES(DATEADD(dd, DATEDIFF(dd, 0, GETDATE()), 0))
Update: did some performance testing, and in this particular case, it seems the amount of work that SQL Server needs to do is really the same - regardless of whether you're using the convert to varchar stripping the time and back to datetime approach that you already have, or whether you're using my get the number of days since date 0 approach. Doesn't seem to make a difference in the end.
The BEST solution however would still be: if you only need the date anyway - use a column of type DATE (in SQL Server 2008 and newer) and save yourself any conversions or manipulations of the GETDATE() output altogether.