How to GROUP BY DATETIME column in SQL - sql

I have a table that tracks when a user scans a logo. It tracks the DATETIME of the scan and also the points awarded to the user for the scan (between 1 and 5)
I am trying to calculate the average points awarded per day for the given week. The challenge I have is when I use GROUP BY I'm getting back every scan because DATETIME value is by the second which can't be grouped.
SET DATEFIRST 1 -- Beginning of week is Monday
SELECT SUM(Points) AS Points, DateTime
FROM SmartTappScanLog
WHERE DateTime >= dateadd(day, 1-datepart(dw, getdate()), CONVERT(date,getdate()))
AND DateTime < dateadd(day, 8-datepart(dw, getdate()), CONVERT(date,getdate()))
GROUP BY DATETIME
I've tried several things including:
GROUP BY CAST(DATETIME As DATE)

You can use DATEPART(), like:
SELECT SUM(Points) AS Points, MIN(DateTime) as DateTime
FROM SmartTappScanLog
WHERE DateTime >= dateadd(day, 1-datepart(dw, getdate()), CONVERT(date,getdate()))
AND DateTime < dateadd(day, 8-datepart(dw, getdate()), CONVERT(date,getdate()))
GROUP BY DATEPART(day,DATETIME)

Related

SQL - How to get record count for each day of week

I have a query that returns all records for the week which starts on Monday. I'm trying to get the record count for each day of the week with no success.
The Original Query:
SET DATEFIRST 1 -- Beginning of week is Monday
SELECT DateTime, ScanPoints, PostPoints
FROM SmartTappScanLog
WHERE DateTime >= dateadd(day, 1-datepart(dw, getdate()), CONVERT(date,getdate()))
AND DateTime < dateadd(day, 8-datepart(dw, getdate()), CONVERT(date,getdate()))
AND UserID = '1' AND BeerID = '3'
ORDER BY DateTime ASC
I've tried several things including Count(DateTime) with a Group By:
SET DATEFIRST 1 -- Beginning of week is Monday
SELECT Count(DateTime), ScanPoints, PostPoints
FROM SmartTappScanLog
WHERE DateTime >= dateadd(day, 1-datepart(dw, getdate()), CONVERT(date,getdate()))
AND DateTime < dateadd(day, 8-datepart(dw, getdate()), CONVERT(date,getdate()))
AND UserID = '1' AND BeerID = '3'
ORDER BY DateTime ASC
GROUP BY DateTime
The error indicates there is a syntax error near GROUP.
Your query has a few issues,
GROUP BY must come before ORDER BY.
Non-aggregated columns selected must also be in the GROUP BY clause
SELECT Count(DateTime), ScanPoints, PostPoints...
GROUP BY ScanPoints, PostPoints
I assume your column DateTime actually contains a date and time. Grouping By this will give you a record for each unique date/time value in your data. Since you're wanting daily data, you should strip off the time by converting it to a date. By the way, DateTime is a data type and a terrible name for a column.
I prefer not to use DATEFIRST as it does not affect all of the date functions equally. If you later use a datediff with weeks, the results may be wrong.
My company also uses Monday as the start of the week. By calculating the shifted first day of the week, you can change which week a Sunday belongs to.
CAST(DATEADD( DAY,
(DATEPART( WEEKDAY, [YourDate] ) + 5) % 7 * -1,
[YourDate] )
AS DATE) AS FirstDayOfWeekIsMonday
Using the FirstDayOfWeekIsMonday column for your week filtering/grouping will ensure that the Sunday is associated with the prior Monday.
As an added bonus, this expression is SARGable so indexes on YourDate can still be used.
Give this a try ( Not tested)
SET DATEFIRST 1 -- Beginning of week is Monday
SELECT datepart(dw, DateTime) as Day, ScanPoints, PostPoints, Count(DateTime) as Recs,
FROM SmartTappScanLog
WHERE DateTime >= dateadd(day, 1-datepart(dw, getdate()), CONVERT(date,getdate()))
AND DateTime < dateadd(day, 8-datepart(dw, getdate()), CONVERT(date,getdate()))
AND UserID = '1' AND BeerID = '3'
GROUP BY datepart(dw, DateTime) , ScanPoints, PostPoints
ORDER BY datepart(dw, DateTime) , ScanPoints, PostPoints
Group by should appear before Order by.
Your Query should be build as below:
SET DATEFIRST 1 -- Beginning of week is Monday
SELECT Count(DateTime), ScanPoints, PostPoints
FROM SmartTappScanLog
WHERE DateTime >=
dateadd(day, 1-datepart(dw, getdate()), CONVERT(date,getdate()))
AND DateTime <
dateadd(day, 8-datepart(dw, getdate()), CONVERT(date,getdate()))
AND UserID = '1' AND BeerID = '3'
GROUP BY DateTime
ORDER BY DateTime ASC

Getdate() functionality returns partial day in select query

I have a query -
SELECT * FROM TABLE WHERE Date >= DATEADD (day, -7, -getdate()) AND Date <= getdate();
This would return all records for each day except day 7. If I ran this query on a Sunday at 17:00 it would only produce results going back to Monday 17:00. How could I include results from Monday 08:00.
Try it like this:
SELECT *
FROM SomeWhere
WHERE [Date] > DATEADD(HOUR,8,DATEADD(DAY, -7, CAST(CAST(GETDATE() AS DATE) AS DATETIME))) --7 days back, 8 o'clock
AND [Date] <= GETDATE(); --now
That's because you are comparing date+time, not only date.
If you want to include all days, you can trunc the time-portion from getdate(): you can accomplish that with a conversion to date:
SELECT * FROM TABLE
WHERE Date >= DATEADD (day, -7, -convert(date, getdate())
AND Date <= convert(date, getdate());
If you want to start from 8 in the morning, the best is to add again 8 hours to getdate.
declare #t datetime = dateadd(HH, 8, convert(datetime, convert(date, getdate())))
SELECT * FROM TABLE
WHERE Date >= DATEADD (day, -7, -#t) AND Date <= #t;
NOTE: with the conversion convert(date, getdate()) you get a datatype date and you cannot add hours directly to it; you must re-convert it to datetime.
Sounds like you want to remove the time. Correct? If so then do the following.
SELECT * FROM TABLE WHERE Date >= (DATEADD (day, -7, -getdate()) AND Date DATEADD(dd, DATEDIFF(dd, 0, getdate()), 0))

How to get week numbers for whole year in SQL Server

My scenario: I have SalesDate (eg: 2007-01-01 to 2016-01-01) filed like this up to todate and have daily date. Now I have to find this salesdate belong to which week.
Condition
Start day of week is Sunday. Based on year, if new year starts week count start as 1.
Try something like below,
DECLARE #SalesDate DATETIME = '2007-01-01'
SELECT DATEDIFF(WEEK, DATEADD(WEEK, DATEDIFF(WEEK, 0, DATEADD(MONTH, DATEDIFF(MONTH, 0, #SalesDate), 0)), 0), #SalesDate - 1) + 1 WeekD, #SalesDate SalesDate
OR
DECLARE #SalesDate DATETIME = '2007-01-01'
SELECT DATEPART(WEEK,#SalesDate) WeekNumber, #SalesDate SalesDate

Only dates from current week

I use the following Query to display the entries in my tables where the date is located in the current week. I need to change it to only show from current date to end of the week, not the hole week.
SET DATEFIRST 1
SELECT distinct Initials
FROM Scheme
WHERE CONVERT(datetime, Dato, 105) >= dateadd(day, 1-datepart(dw, getdate()), CONVERT(date,getdate()))
AND convert(datetime, Dato, 105) < dateadd(day, 8-datepart(dw, getdate()), CONVERT(date,getdate()))
AND RoomId = ? AND Initials IS NOT NULL
A bit different, but more effective method:
SELECT distinct Initials
FROM Scheme
WHERE Dato >= cast(current_timestamp as date) -- currentdate
AND Dato < dateadd(d, datediff(d, -7, current_timestamp)/7*7, 0)--end of week
AND RoomId = ? AND Initials IS NOT NULL
I assume Dato is defined as a date or datetime. Otherwise you may encounter some performance issues and this will not work.
dateadd(day, 1-datepart(dw, getdate()), CONVERT(date,getdate()))
This means: add 1-datepart(dw, getdate()) days to current date.
This means 1-datepart(dw, getdate()): go to starting date of current week.
But you need just current date CONVERT(date,getdate()).
Just change your code to:
SET DATEFIRST 1
SELECT distinct Initials
FROM Scheme
WHERE CONVERT(datetime, Dato, 105) >= CONVERT(date,getdate())
AND convert(datetime, Dato, 105) < dateadd(day, 8-datepart(dw, getdate()), CONVERT(date,getdate()))
AND RoomId = ? AND Initials IS NOT NULL

Get the records of last month in SQL server

I want to get the records of last month based on my db table [member] field "date_created".
What's the sql to do this?
For clarification,
last month - 1/8/2009 to 31/8/2009
If today is 3/1/2010, I'll need to get the records of 1/12/2009 to 31/12/2009.
All the existing (working) answers have one of two problems:
They will ignore indices on the column being searched
The will (potentially) select data that is not intended, silently corrupting your results.
1. Ignored Indices:
For the most part, when a column being searched has a function called on it (including implicitly, like for CAST), the optimizer must ignore indices on the column and search through every record. Here's a quick example:
We're dealing with timestamps, and most RDBMSs tend to store this information as an increasing value of some sort, usually a long or BIGINTEGER count of milli-/nanoseconds. The current time thus looks/is stored like this:
1402401635000000 -- 2014-06-10 12:00:35.000000 GMT
You don't see the 'Year' value ('2014') in there, do you? In fact, there's a fair bit of complicated math to translate back and forth. So if you call any of the extraction/date part functions on the searched column, the server has to perform all that math just to figure out if you can include it in the results. On small tables this isn't an issue, but as the percentage of rows selected decreases this becomes a larger and larger drain. Then in this case, you're doing it a second time for asking about MONTH... well, you get the picture.
2. Unintended data:
Depending on the particular version of SQL Server, and column datatypes, using BETWEEN (or similar inclusive upper-bound ranges: <=) can result in the wrong data being selected. Essentially, you potentially end up including data from midnight of the "next" day, or excluding some portion of the "current" day's records.
What you should be doing:
So we need a way that's safe for our data, and will use indices (if viable). The correct way is then of the form:
WHERE date_created >= #startOfPreviousMonth AND date_created < #startOfCurrentMonth
Given that there's only one month, #startOfPreviousMonth can be easily substituted for/derived by:
DATEADD(month, -1, #startOfCurrentMonth)
If you need to derive the start-of-current-month in the server, you can do it via the following:
DATEADD(month, DATEDIFF(month, 0, CURRENT_TIMESTAMP), 0)
A quick word of explanation here. The initial DATEDIFF(...) will get the difference between the start of the current era (0001-01-01 - AD, CE, whatever), essentially returning a large integer. This is the count of months to the start of the current month. We then add this number to the start of the era, which is at the start of the given month.
So your full script could/should look similar to the following:
DECLARE #startOfCurrentMonth DATETIME
SET #startOfCurrentMonth = DATEADD(month, DATEDIFF(month, 0, CURRENT_TIMESTAMP), 0)
SELECT *
FROM Member
WHERE date_created >= DATEADD(month, -1, #startOfCurrentMonth)
AND date_created < #startOfCurrentMonth
All date operations are thus only performed once, on one value; the optimizer is free to use indices, and no incorrect data will be included.
SELECT *
FROM Member
WHERE DATEPART(m, date_created) = DATEPART(m, DATEADD(m, -1, getdate()))
AND DATEPART(yyyy, date_created) = DATEPART(yyyy, DATEADD(m, -1, getdate()))
You need to check the month and year.
Add the options which have been provided so far won't use your indexes at all.
Something like this will do the trick, and make use of an index on the table (if one exists).
DECLARE #StartDate DATETIME, #EndDate DATETIME
SET #StartDate = dateadd(mm, -1, getdate())
SET #StartDate = dateadd(dd, datepart(dd, getdate())*-1, #StartDate)
SET #EndDate = dateadd(mm, 1, #StartDate)
SELECT *
FROM Member
WHERE date_created BETWEEN #StartDate AND #EndDate
DECLARE #StartDate DATETIME, #EndDate DATETIME
SET #StartDate = DATEADD(mm, DATEDIFF(mm,0,getdate())-1, 0)
SET #EndDate = DATEADD(mm, 1, #StartDate)
SELECT *
FROM Member
WHERE date_created BETWEEN #StartDate AND #EndDate
An upgrade to mrdenny's solution, this way you get exactly last month from YYYY-MM-01
Last month consider as till last day of the month.
31/01/2016 here last day of the month would be 31 Jan. which is not similar to last 30 days.
SELECT CONVERT(DATE, DATEADD(DAY,-DAY(GETDATE()),GETDATE()))
One way to do it is using the DATEPART function:
select field1, field2, fieldN from TABLE where DATEPART(month, date_created) = 4
and DATEPART(year, date_created) = 2009
will return all dates in april. For last month (ie, previous to current month) you can use GETDATE and DATEADD as well:
select field1, field2, fieldN from TABLE where DATEPART(month, date_created)
= (DATEPART(month, GETDATE()) - 1) and
DATEPART(year, date_created) = DATEPART(year, DATEADD(m, -1, GETDATE()))
declare #PrevMonth as nvarchar(256)
SELECT #PrevMonth = DateName( month,DATEADD(mm, DATEDIFF(mm, 0, getdate()) - 1, 0)) +
'-' + substring(DateName( Year, getDate() ) ,3,4)
SQL query to get record of the present month only
SELECT * FROM CUSTOMER
WHERE MONTH(DATE) = MONTH(CURRENT_TIMESTAMP) AND YEAR(DATE) = YEAR(CURRENT_TIMESTAMP);
SELECT * FROM Member WHERE month(date_created) = month(NOW() - INTERVAL 1 MONTH);
select * from [member] where DatePart("m", date_created) = DatePart("m", DateAdd("m", -1, getdate())) AND DatePart("yyyy", date_created) = DatePart("yyyy", DateAdd("m", -1, getdate()))
DECLARE #StartDate DATETIME, #EndDate DATETIME
SET #StartDate = DATEADD(mm, DATEDIFF(mm, 0, getdate()) - 1, 0)
SET #EndDate = dateadd(dd, -1, DATEADD(mm, 1, #StartDate))
SELECT * FROM Member WHERE date_created BETWEEN #StartDate AND #EndDate
and another upgrade to mrdenny's solution.
It gives the exact last day of the previous month as well.
WHERE
date_created >= DATEADD(MONTH, DATEDIFF(MONTH, 31, CURRENT_TIMESTAMP), 0)
AND date_created < DATEADD(MONTH, DATEDIFF(MONTH, 0, CURRENT_TIMESTAMP), 0)
I'm from Oracle env and I would do it like this in Oracle:
select * from table
where trunc(somedatefield, 'MONTH') =
trunc(sysdate -INTERVAL '0-1' YEAR TO MONTH, 'MONTH')
Idea: I'm running a scheduled report of previous month (from day 1 to the last day of the month, not windowed). This could be index unfriendly, but Oracle has fast date handling anyways.
Is there a similar simple and short way in MS SQL? The answer comparing year and month separately seems silly to Oracle folks.
You can get the last month records with this query
SELECT * FROM dbo.member d
WHERE CONVERT(DATE, date_created,101)>=CONVERT(DATE,DATEADD(m, datediff(m, 0, current_timestamp)-1, 0))
and CONVERT(DATE, date_created,101) < CONVERT(DATE, DATEADD(m, datediff(m, 0, current_timestamp)-1, 0),101)
I don't think the accepted solution is very index friendly
I use the following lines instead
select * from dbtable where the_date >= convert(varchar(10),DATEADD(m, -1, dateadd(d, - datepart(dd, GETDATE())+1, GETDATE())),120) and the_date <= dateadd(ms, -3, convert(varchar(10),DATEADD(m, 0, dateadd(d, - datepart(dd, GETDATE())+1, GETDATE())),120));
Or simply (this is the best).
select * from dbtable where the_date >= convert(varchar(10),DATEADD(m, -1, dateadd(d, - datepart(dd, GETDATE())+1, GETDATE())),120) and the_date < SELECT convert(varchar(10),DATEADD(m, -1, dateadd(d, - datepart(dd, GETDATE())+1, GETDATE())),120);
Some help
-- Get the first of last month
SELECT convert(varchar(10),DATEADD(m, -1, dateadd(d, - datepart(dd, GETDATE())+1, GETDATE())),120);
-- Get the first of current month
SELECT convert(varchar(10),DATEADD(m, -1, dateadd(d, - datepart(dd, GETDATE())+1, GETDATE())),120);
--Get the last of last month except the last 3milli seconds. (3miliseconds being used as SQL express otherwise round it up to the full second (SERIUSLY MS)
SELECT dateadd(ms, -3, convert(varchar(10),DATEADD(m, 0, dateadd(d, - datepart(dd, GETDATE())+1, GETDATE())),120));
Here is what I did so I could put it in a view:
ALTER view [dbo].[MyView] as
with justdate as (
select getdate() as rightnow
)
, inputs as (
select dateadd(day, 1, EOMONTH(jd.rightnow, -2)) as FirstOfLastMonth
,dateadd(day, 1, EOMONTH(jd.rightnow, -1)) as FirstOfThisMonth
from justdate jd
)
SELECT TOP 10000
[SomeColumn]
,[CreatedTime]
from inputs i
join [dbo].[SomeTable]
on createdtime >= i.FirstOfLastMonth
and createdtime < i.FirstOfThisMonth
order by createdtime
;
Note that I intentionally ran getdate() once.
In Sql server for last one month:
select * from tablename
where order_date > DateAdd(WEEK, -1, GETDATE()+1) and order_date<=GETDATE()
DECLARE #curDate INT = datepart( Month,GETDATE())
IF (#curDate = 1)
BEGIN
select * from Featured_Deal
where datepart( Month,Created_Date)=12 AND datepart(Year,Created_Date) = (datepart(Year,GETDATE())-1)
END
ELSE
BEGIN
select * from Featured_Deal
where datepart( Month,Created_Date)=(datepart( Month,GETDATE())-1) AND datepart(Year,Created_Date) = datepart(Year,GETDATE())
END
DECLARE #StartDate DATETIME, #EndDate DATETIME
SET #StartDate = dateadd(mm, -1, getdate())
SET #StartDate = dateadd(dd, datepart(dd, getdate())*-1, #StartDate)
SET #EndDate = dateadd(mm, 1, #StartDate)
set #StartDate = DATEADD(dd, 1 , #StartDate)
The way I fixed similar issue was by adding Month to my SELECT portion
Month DATEADD(day,Created_Date,'1971/12/31') As Month
and than I added WHERE statement
Month DATEADD(day,Created_Date,'1971/12/31') = month(getdate())-1
If you are looking for last month so try this,
SELECT
FROM #emp
WHERE DATEDIFF(MONTH,CREATEDDATE,GETDATE()) = 1
If you are looking for last month so try this,
SELECT
FROM #emp
WHERE DATEDIFF(day,CREATEDDATE,GETDATE()) between 1 and 30
A simple query which works for me is:
select * from table where DATEADD(month, 1,DATEFIELD) >= getdate()
If you are looking for previous month data:
date(date_created)>=date_sub(date_format(curdate(),"%Y-%m-01"),interval 1 month) and
date(date_created)<=date_sub(date_format(curdate(),'%Y-%m-01'),interval 1 day)
This will also work when the year changes. It will also work on MySQL.