I'm executing the next query in sql server 2012.
select *
from table
where date > convert(date, '2015/02/12')
order by date asc
but I'm getting the next set:
2015-02-12 06:40:42.000
2015-02-12 06:45:44.000
2015-02-12 06:48:15.000
2015-02-12 07:06:28.000
2015-02-12 07:26:46.000
...
I can fix this by changing the date to '2015/02/13', but I have the doubt about this behavior, why am I getting dates from feb 12 when I am specifying that I need only later dates ?. I also tried using cast('2015/02/12' as date), but I could not have the answer I was looking for
Because dates without times are intepreted as 12:00 midnight on that date. If you want only dates after February 12, 2015, then select all dates greater than or equal to February 13, 2015 (which again will be interpreted as midnight on February 13th).
select *
from table
where date >= convert(date, '2015/02/13')
order by date asc
why am I getting dates from feb 12 when I am specifying that I need only later dates ?
You're not specifying that you need only dates after Feb 12. You're asking for every row in which the value in the "date" column is greater than '2015-02-12'.
The value '2015-02-12 06:40:42.000' is greater than '2015-02-12'.
When comparing the date 2015/02/12 with your datetime data, this will implicitly compare the converted date 2015-02-12 00:00:000, the date at the beginning of the day with all of your data in column date.
But you are actually comparing datetime data, which has a time part as well, which gets compared.
And because you're comparing the beginning of the day (2015-02-12 00:00:000) with a value which is after it, for example 2015-02-12 06:40:42, all of the dates from will be displayed, because 6:40 AM is after (greater than) 0:00 AM.
Try this:
SELECT *
FROM TABLE
WHERE DATE >= DATEADD(SECOND, -1, '2015/02/13')
jarlh is right, though I'll clarify a little. Each of the "dates" you show above fall after 12:00 midnight starting 2015-02-12. They are actually timestamps.
If you don't want to see anything for the day specified in the filter, you add a day and use the greater-than-or-equal-to (>=) operator.
SELECT *
FROM table
WHERE (date >= DATEADD(d, 1, CONVERT(date, '2015/02/12')))
ORDER BY date ASC
Related
Using below depart syntax to fetch for previous month record, it is working fine till previous year,however it is giving 0 value in January month.How can we get pervious month with date part even if year is change ?
DATEPART(month(GETDATE()) -1
I understand that I used another type of DB, but I want to give a hint. I am using sql server 2019.
Firstly, you need to substitute date and only then take datepart from it.
Queries:
--dateadd -1 would subtract 1 from current month
--(Jan - 1 2022), would be December 2021
select datepart(month, dateadd(month, -1, getdate()))
--also date add covers internally the problem with 30,31 days.
--May always has 31 days, April 30. So -1 subtraction from 31th of May,would result in 30th of April.
select dateadd(month, -1, cast('2021-05-31 10:00:00' as datetime))
I want to select a date that gives me 18 months from the current date, but I am not getting exactly 18 months.
Before Running the query, min and max dates are.
MIN Date: 02/22/2018
MAX Date: 03/26/2021
Query
SELECT DATEADD(dd,-548,cast(GETDATE() as datetime));
Above query returns the min date as 10/15/2019. So from 10/15/2019 to 03/26/2021 is 17 months and 12 days. While my expectation is to get 10/05/2019 from this query so that I get exactly 18 months or 548 days.
SELECT DATEADD(month,-18,cast(GETDATE() as datetime))
The reason I wasn't getting a correct date was because the query looks for current date GETDATE() and in the database, I don't have data for the current date due to which I was encountering differences. I exported data from STAGE to DEV environment and now I am getting correct date.
I want to retrieve the data between a 6 day time period.
The output I want is:
Date
--------
2019-05-01
2019-05-04
2019-06-01
2019-06-06
2019-07-01
This is my query so far:
select date from data d
where CAST(d.createdate as Date) between CAST('2019-05-01' as Date)
AND DATEADD(CAST(dd,6,'2016-07-01') as Date)
Why is this not retrieving the results I want?
You have several problems with your query.
The first is with your DATEADD statement which is all mixed up. You are not nesting the casted date into the statement properly. This is the corrected version:
DATEADD(dd, 6, CAST('2016-07-01' as Date))
The second is that your select projection refers to the column date which does not exist. Instead, you probably want your createdate column.
The third is that your between clause is back to front. You are saying between 2019-05-01 and 2016-07-01 but the smaller date must come first.
In fact, your given example is incorrect. In your question, you say "want to retrieve the data between two dates only for 6 days." So, why would you start with a date in 2016 and then jump to a date in 2019 and add 6 days to the date in 2019? If you want to use the DATEADD approach, you need to use the same date in both positions.
So here is your corrected query:
select d.createdate from data d
where CAST(d.createdate as Date) between CAST('2019-05-01' as Date)
AND DATEADD(dd, 6, CAST('2019-05-01' as Date))
I have a large data file containing a string type column 'YearMonthWeek'
It contains values such as '20160101' for the first week of January 2016, or '20161040' for the 40th week of the year 2016 apparently falling in October.
Now, I want to convert these strings to actual dates so that every YearMonthWeek value is converted to, say the first day of that week. (Whether that ends up being Monday or Sunday I don't really care).
I tried the following query:
PARSE_TIMESTAMP('%Y%m%W', CAST(YearMonthWeek AS STRING)) AS datefield
(See this documentation for details)
This runs without errors, but returns me the first day of the month for every single entry...
So for example '20160101' and '20160102' both get parsed as 2016-01-01 00:00:00 UTC.
Is this an issue with the PARSE_TIMESTAMP function, or am I missing something?
Try doing something like
DATE_ADD(date_expression, INTERVAL %W WEEK)
Static example:
SELECT
DATE_ADD(
DATE(PARSE_TIMESTAMP('%Y', SUBSTR(CAST('20161252' AS STRING),0,4))),
INTERVAL (CAST(SUBSTR(CAST('20160102' AS STRING),7) AS INT64)) week)
AS datefield
-
Row datefield
1 2016-01-15
You may add something as a margin to it, according to ISO 8601, the first week of the year is the one that contains January 4th. So you could have something like: 4 + 7*($week - 1)
I am trying to get the last of month, and in order to that i have written the following, to calculate the no. of days between today and the last date.
select datediff(DAY,GETDATE(),dateadd(m,1,getdate()))-GETDATE()
the bold part gives me the no. of days between today and a month from today, say 30 or 31. and then I am subtracting today's date from 30 or 31, which is " -getdate() "
The output for the above query is
1786-06-06 11:44:30.540
Could you please explain what is happening in the query? I am not looking for a solution, I would like to know how is SQL-Server interpreting the query.
Thanks. :)
The bold part of the expressions does not return a date, it returns a number of days:
31
Convert that to a datetime:
SELECT CONVERT(DATETIME, 31);
This is 31 days after day 0 (1900-01-01):
1900-02-01
Now, subtract GETDATE() as an integer (41512 days after day 0):
SELECT 31 - 41512 = -41481
Now add -41481 days to day 0:
SELECT DATEADD(DAY, -41481, 0);
-- or
SELECT DATEADD(DAY, -41481, '19000101');
Or:
SELECT CONVERT(DATETIME, 31 - CONVERT(INT, GETDATE()));
Now, I strongly recommend a couple of things:
Don't use implicit date math. #date_var_or_col - 1 for example fails with new data types like DATE and DATETIME2.
Don't use shorthand like m. If you mean MONTH, just take the massive productivity hit and type out MONTH. To see why, tell me if this provides the results you expect:
SELECT DATEPART(y, GETDATE()), DATEPART(w, GETDATE());
I am subtracting today's date from 30 or 31, which is " -getdate() "
Sounds like you understand exactly what is happening, but maybe don't understand the results.
You are implicitly converting GETDATE() to a number, which represents the number of days (and fractional days) since 1/1/1900 12:00:00 AM
When you "subtract" GETDATE() (41,511 as of 8/27/2013) from 30 or 31 you get an answer of -41,480, or 41,480 days before 1/1/1900, which would be about 6/6/1786 (plus or minus a few hours for the fractional part).