Have been searching for the answer to this for a which but with no joy. Hoping you DAX geniuses can help out!
I have a table of transactional data with a date time column (in the format "dd/mm/yyyy hh:mm:ss")
I want to look this datetime up in a separate 'shift reference' table to add a new column to my transactional data i.e. if it falls between 2 date times (which it always will), Start time and End Time there will be a corresponding shift associated with it.
The format of this table is
Start time - End Time - Shift Pattern
In this table we have the datetime (in the same format as before) the shift started - "Start_Time", when it ended - "End_Time" and what 'Shift' was working. I want to use my transactional datetime to look up what shift was on when the transaction took place.
Ive tried combinations of Lookupvalue/Calculate/Max and on some occasions it has returned values, but never correct ones!
I hope this makes sense!
Best Regards,
Colin
You can use this code to add a calculated column with the Shift value looked up based on transaction timestamp.
Shift = CALCULATE (
VALUES ( Shift[SHIFT_PATTERN] ),
FILTER (
Shift,
Shift[start_time] <= [Timestamp] && Shift[end_time] > [Timestamp]
)
)
Just for showing another option, it is even possible to calculate transaction facts sliced by Shift without adding a column in the fact table. You may handle it in measures like below.
Transaction Count =
SUMX (
Shift,
COUNTROWS (
FILTER (
'Transaction',
'Transaction'[Timestamp] >= Shift[start_time]
&& 'Transaction'[Timestamp] < Shift[end_time]
)
)
)
I have this database
Table [Ticks]
Fields: 2
[Value]: INT
[Time]: DATETIME
Indexes: 1
[TicksIdx]
[Time]
[Value]
Triggers: 0
Table constraints:
Primary Key:
Fields:
On Conflict:
Foreign Keys: 0
Unique constraints: 0
Check constraints: 0
Table [Ticks] end
And I want to count all the records with value=-1 and Time < (Now-10 hours) from 10 hours ago until now
I have this query and always returns 0
select count(*) from Ticks where Value=-1 and
time>=datetime('now', '-10 hours')
If I change the query to this then it returns all the records
select count(*) from Ticks where Value=-1 and
time<datetime('now', '-10 hours')
and I have records with time values more recent than 10 hours
Explanations
You issue will be due to the value stored in the time column not being in a recognised format, complicated also with not comparing like for like.
The SQLite recognised formats can be found here SQL As Understood By SQLite - Date And Time Functions - Time Strings.
For example YYYY-MM-DD HH:MM:SS (e.g. 2018-12-09 12:40:01) is such a recognised format.
You query is (assuming 9/12/2018 12:40:01 AM from c# and the runtime is the same) saying:-
is 9/12/2018 12:40:01 AM >= 2018-12-09 12:40:01
as the values are non-numeric then a character by character comparison is undertaken so using these times 9 IS greater than 2 (of course if it were 09/12..... then IS NOT).
The correct way, to ensure like for like comparison, if the underlying data is in a recognised format, would be to ensure that both sides compare dimetime's as such really the SQL should therefore be :-
SELECT count(*) FROM Ticks WHERE Value=-1 AND datetime(time) >= datetime('now', '-10 hours')
note datetime(time) may only be needed if the date in the time column were another recognised format than the example format.
-note you may need to incorporate localtime to adjust times e.g. datetime(time) >= datetime('now','localtime', '-10 hours')
as per
The "localtime" modifier (12) assumes the time string to its left is
in Universal Coordinated Time (UTC) and adjusts the time string so
that it displays localtime. If "localtime" follows a time that is not
UTC, then the behavior is undefined. The "utc" modifier is the
opposite of "localtime". "utc" assumes that the string to its left is
in the local timezone and adjusts that string to be in UTC. If the
prior string is not in localtime, then the result of "utc" is
undefined.
Fix(es)
To fix the issue you would need to either
do a complex reformat of the time column, dependant upon the format of the value in the time column, for the comparison (not recommended)
or ensure that the value is stored in a SQlite recognised format.
If the time is the current time when the ticks row is inserted, then you could use time DATETIME DEFAULT CURRENT_TIMESTAMP as the definition for the time column and not supply a value for the time column when inserting a row.
Otherwise you should change the format of the value to a recognised format before storing the value.
I have a table that includes a 'LastUpdated' column that is generated when the row is inserted using Sqlite's datetime('now') function.
How do I write a Select statement that finds all rows with 'LastUpdated' more than 100 days old?
I think it's a variant of:
SELECT * FROM Table WHERE (DATETIME('Now')-100 Days) > LastUpdated
But I'm unsure of:
a) How to specify the 100 Days?
b) Whether I can actually compare datetimes like this or if I first have to convert DATETIME('Now') to a string?
c) DATETIME('Now') results in UTC time, correct? I think so from my reading of the documentation, but it was a little confusing...
Figured it out--I didn't see all the handy modifiers at the bottom of the SQLite Datetime Documentation.
A bunch of helpful examples there demonstrating addition/subtraction of any datetime unit (years, months, hours, seconds, etc)
SELECT * FROM Table WHERE (DATETIME('Now','-100 Days') > LastUpdated
I'm creating a report from a stored procedure that pulls two date/times (CreatedDate and ClosedDate). I need a column on the report that shows the difference (i.e. time it took to go from open to close). First, I just subtracted CreatedDate from ClosedDate (in the report [SQL Server Reporting Services], not in the stored procedure) and got a time that looks like this: 72.20:34:18.6230000 (day.hour:minute:second). I need to shrink this down, if possible, to just day.hour...
I was experimenting with some of the functions found on MSDN (http://msdn.microsoft.com/en-us/library/ms186724.aspx). DATEDIFF almost gives me what I need, but I can only specify days or hours, and ideally (as I said), I need it to show the 'time to close' as both (day.hour).
Is this possible?
In SSRS, you can apply a custom format to that column to show only days.hours. Right click the column in design mode -->text box properties-->Number(on left hand side). If you don't see one of the formats for date, time or number that fits what you need, create a custom one at the bottom.
Convert to minutes
divide by 1440 gives whole days
modulo 1440 gives remaining minutes, divide by 60 for hours
Something like (not tested):
SELECT
CAST(DATEDIFF(minute, CreatedDate, ClosedDate) / 1440 AS varchar(20)) + '.'
CAST((DATEDIFF(minute, CreatedDate, ClosedDate) % 1440) / 60 AS varchar(20))
FROM
MyTable
You may need to fiddle with the hours representation in cade I've misunderstood
DATEDIFF for day and hour go by boundaries: that is if there are just 3 minutes between the 2 values spanning midnight, there will be one hour/day difference. So I used minutes
Edit:
To overflow the int from DATEDIFF requires a difference of 4000+ years
Thoughts:
Using datetime2 and have CreatedDate of lowest 0001-01-01
ClosedDate is a sentinel value like 9999-12-31 say for "open" items
CreatedDate and ClosedDate are varchar and conversion to datetime is faulty
Your example shows 72 days difference which would be around 104k minutes.
I would try this to see where you have more than 1000 year differences which would be a mere half billion or so minutes:
SELECT * FROM mytables
WHERE DATEDIFF(year, CreatedDate, ClosedDate) > 1000
How do I get the difference in days between 2 dates in SQLite? I have already tried something like this:
SELECT Date('now') - DateCreated FROM Payment
It returns 0 every time.
SELECT julianday('now') - julianday(DateCreated) FROM Payment;
Difference In Days
Select Cast ((
JulianDay(ToDate) - JulianDay(FromDate)
) As Integer)
Difference In Hours
Select Cast ((
JulianDay(ToDate) - JulianDay(FromDate)
) * 24 As Integer)
Difference In Minutes
Select Cast ((
JulianDay(ToDate) - JulianDay(FromDate)
) * 24 * 60 As Integer)
Difference In Seconds
Select Cast ((
JulianDay(ToDate) - JulianDay(FromDate)
) * 24 * 60 * 60 As Integer)
Both answers provide solutions a bit more complex, as they
need to be. Say the payment was created on January 6, 2013.
And we want to know the difference between this date and today.
sqlite> SELECT julianday() - julianday('2013-01-06');
34.7978485878557
The difference is 34 days. We can use julianday('now') for
better clarity. In other words, we do not need to put
date() or datetime() functions as parameters to julianday()
function.
The SQLite documentation is a great reference and the DateAndTimeFunctions page is a good one to bookmark.
It's also helpful to remember that it's pretty easy to play with queries with the sqlite command line utility:
sqlite> select julianday(datetime('now'));
2454788.09219907
sqlite> select datetime(julianday(datetime('now')));
2008-11-17 14:13:55
This answer is a little long-winded, and the documentation will not tell you this (because they assume you are storing your dates as UTC dates in the database), but the answer to this question depends largely on the timezone that your dates are stored in. You also don't use Date('now'), but use the julianday() function, to calculate both dates back against a common date, then subtract the difference of those results from each other.
If your dates are stored in UTC:
SELECT julianday('now') - julianday(DateCreated) FROM Payment;
This is what the top-ranked answer has, and is also in the documentation. It is only part of the picture, and a very simplistic answer, if you ask me.
If your dates are stored in local time, using the above code will make your answer WRONG by the number of hours your GMT offset is. If you are in the Eastern U.S. like me, which is GMT -5, your result will have 5 hours added onto it. And if you try making DateCreated conform to UTC because julianday('now') goes against a GMT date:
SELECT julianday('now') - julianday(DateCreated, 'utc') FROM Payment;
This has a bug where it will add an hour for a DateCreated that is during Daylight Savings Time (March-November). Say that "now" is at noon on a non-DST day, and you created something back in June (during DST) at noon, your result will give 1 hour apart, instead of 0 hours, for the hours portion. You'd have to write a function in your application's code that is displaying the result to modify the result and subtract an hour from DST dates. I did that, until I realized there's a better solution to that problem that I was having: SQLite vs. Oracle - Calculating date differences - hours
Instead, as was pointed out to me, for dates stored in local time, make both match to local time:
SELECT julianday('now', 'localtime') - julianday(DateCreated) FROM Payment;
Or append 'Z' to local time:
julianday(datetime('now', 'localtime')||'Z') - julianday(CREATED_DATE||'Z')
Both of these seem to compensate and do not add the extra hour for DST dates and do straight subtraction - so that item created at noon on a DST day, when checking at noon on a non-DST day, will not get an extra hour when performing the calculation.
And while I recognize most will say don't store dates in local time in your database, and to store them in UTC so you don't run into this, well not every application has a world-wide audience, and not every programmer wants to go through the conversion of EVERY date in their system to UTC and back again every time they do a GET or SET in the database and deal with figuring out if something is local or in UTC.
Just a note for writing timeclock functions. For those looking for hours worked, a very simple change of this gets the hours plus the minutes are shown as a percentage of 60 as most payroll companies want it.
CAST ((julianday(clockOUT) - julianday(clockIN)) * 24 AS REAL) AS HoursWorked
Clock In Clock Out HoursWorked
2016-08-07 11:56 2016-08-07 18:46 6.83333332836628
Given that your date format follows : "YYYY-MM-DD HH:MM:SS",
if you need to find the difference between two dates in number of months :
(strftime('%m', date1) + 12*strftime('%Y', date1)) -
(strftime('%m', date2) + 12*strftime('%Y', date2))
Firstly, it's not clear what your date format is.
There already is an answer involving strftime("%s").
I like to expand on that answer.
SQLite has only the following storage classes: NULL, INTEGER, REAL, TEXT or BLOB.
To simplify things, I'm going to assume dates are REAL containing the seconds since 1970-01-01.
Here's a sample schema for which I will put in the sample data of "1st December 2018":
CREATE TABLE Payment (DateCreated REAL);
INSERT INTO Payment VALUES (strftime("%s", "2018-12-01"));
Now let's work out the date difference between "1st December 2018" and now (as I write this, it is midday 12th December 2018):
Date difference in days:
SELECT (strftime("%s", "now") - DateCreated) / 86400.0 FROM Payment;
-- Output: 11.066875
Date difference in hours:
SELECT (strftime("%s", "now") - DateCreated) / 3600.0 FROM Payment;
-- Output: 265.606388888889
Date difference in minutes:
SELECT (strftime("%s", "now") - DateCreated) / 60.0 FROM Payment;
-- Output: 15936.4833333333
Date difference in seconds:
SELECT (strftime("%s", "now") - DateCreated) FROM Payment;
-- Output: 956195.0
If you want time in 00:00 format:
I solved it like that:
SELECT strftime('%H:%M',
CAST((julianday(FinishTime) - julianday(StartTime)) AS REAL),
'12:00')
FROM something;
If you want difference in seconds
SELECT strftime('%s', '2019-12-02 12:32:53') - strftime('%s', '2019-12-02 11:32:53')
If you want records in between days,
select count(col_Name) from dataset where cast(julianday("now")- julianday(_Last_updated) as int)<=0;
In my case, I have to calculate the difference in minutes and julianday() does not give an accurate value. Instead, I use strftime():
SELECT (strftime('%s', [UserEnd]) - strftime('%s', [UserStart])) / 60
Both dates are converted to unixtime (seconds), then subtracted to get
value in seconds between the two dates. Next, divide it by 60.
https://www.sqlite.org/cvstrac/wiki?p=DateAndTimeFunctions