I am doing this
datediff = (date1 - DateOfDeposit).TotalDays;
But this gives the no. of days and I want no.of months.
The simplest option is probably to use my Noda Time library, which was designed for exactly this sort of thing (as well as making it cleaner to work with dates and times in general):
LocalDate start = new LocalDate(2013, 1, 5);
LocalDate end = new LocalDate(2014, 6, 1);
Period period = Period.Between(start, end, PeriodUnits.Months);
Console.WriteLine(period.Months); // 16
There's nothing built into .NET to make this particularly easy. You could subtract years and months as per Ani's answer, but then also take the day of month into account to avoid the issue I describe in comments. I'd suggest writing a good set of unit tests first though - bearing in mind leap years and the like.
Assuming you want the number of whole months, I think this should work (not sure what corner-cases this doesn't handle, but can't think of any off the top of my head):
12 * (date1.Year - DateOfDeposit.Year) + (date1.Month - DateOfDeposit.Month)
You could always add any fractional component with DateTime.Day, but it's not clear how that should precisely be defined (how many months are between Jan 29 and Feb 27?).
If you're doing a lot of date-time handling, Jon Skeet's Noda Time library (that he mentions in his answer) is a good choice.
Related
Kusto provides functions to get the start- and end-day of the week. These are invoked through startofweek() and endofweek(). startofweek() returns Sunday, and endofweek() returns Saturday.
In some countries, weeks are from Monday to Sunday, which I have been unable to locate how to change.
Is it possible to set the culture in kusto, such that startofweek() and endofweek() would return Monday and Sunday respectively?
No, I have not seen a setting where culture can be set in Kusto. Where operating systems and applications which make it possible to set such configurations, I see Kusto being being fairly basic and similar to .NET libraries where the DayOfWeek enum begins with Sunday at index 0 and there is no way to change that foundation.
The documentation indicates that "Start of the week is considered to be a Sunday." so I believe that is that. It is up to us users to adapt from there. (https://learn.microsoft.com/en-us/azure/data-explorer/kusto/query/startofweekfunction)
I'm sure you have done this yourself already, but here has been my approach.
startofweek(now()) + 1d
endofweek(now()) + 1d
My boss asked me to perform a T-Test to test the significance for a certain metric we use called conversion rate.
I have collected 18 months worth of data for this metric dating April 1, 2017 - September 30th, 2018.
He initially told me to collect 12 - 14 months of the data and run a t-test to to look for significance of the metric. (Higher conversion rate means better!).
I'm not really sure how to go about it. Do I split the data up into 9 month samples i.e. Sample 1: April 2017 - December 2017, Sample 2: January 2018 - September 2018 and run a two sample t-test? Or would it make sense to compare all of the data against a mean like 0?
Is there a better approach to this? The bottom line is he wants to see that the conversion rate has significantly increased over time.
Thanks,
- Keith
My advice is to dump the t-test and look only at the magnitude of the change in the conversion rate. After all, the conversion rate is what's important to your business. By the way, looking at the magnitude of something practically relevant is called "effect size analysis"; a web search for that should turn up a lot of resources. To get started, just make a plot of the available data -- is conversion rate going up or going down or what?
Further questions should be directed to stats.stackexchange.com instead of SO. Good luck and have fun.
I am playing around with NodeMCU on an ESP8266. I have a Date String and a Time String from a Web Request like this:
15.07.16 (German format DD.MM.YY)
19:50 (24 hours format)
These DateTimes lay usually a little bit in the future. I want to get the number of minutes from the current time to the time from my strings above.
I guess I have to create a time object from the strings and then compare it to the current time. But how can I do that with Lua?
Unfortunately there is no os Library on NodeMCU (or I might have missed how to enable it).
Calculating the difference manually would be a huge pain which I would like to avoid. Does anyone know a way to compute that with available or external libraries?
Thanks for any support!
There's a pending PR for rtctime that does the exact opposite, Unix epoch to UTC calendar.
If you convert your strings to a Unix epoch X you could do
-- delta in minutes
local delta = (X - rtctime.get()) / 60
You can either calculate X yourself, which is far from trivial due to leap years & seconds and other date/time oddities, or your can send a request to http://www.convert-unix-time.com/api?date=15.07.2016%2019:50&timezone=Vienna&format=german and extract the timestamp from it.
First you get the numbers from the strings using Lua's string library:
https://www.lua.org/pil/20.html
https://www.lua.org/manual/5.3/manual.html#6.4
Then you do the time calculations using Lua's os library:
https://www.lua.org/pil/22.1.html
https://www.lua.org/manual/5.3/manual.html#6.9
I won't give you more information as you did not show any own effort to solve the problem.
Addon:
As you don't have the os library (didn't know that) you can simply calculate that stuff yourself.
Get the month, year hour and minute number from the strings using string.sub or string patterns.
Then simply calculate the time difference. You know how many days each month has. You know how many minutes per hour and how many hours per day.
Determine if the year is a leap year (if you don't know how: https://support.microsoft.com/en-us/kb/214019)
I thought of Integer 16, because it's easy to calculate years, depending from it.
String is good for searching, for ex. it's easy to find year 78 as substring from 1978.
And it's also possible to keep year as Date.
What's the best practice?
I would think that storing something that involves a date of some kind would be best stored as an NSDate simply because it's pretty likely that when you pull that value back out, it will be for displaying some kind of date. I can't imagine that XCode would go out of the way to make NSDate so accessible in Core Data unless they had a good reason.
It could also be optimization...
Lastly, it's all about the circumstance. Like you had stated: yes, sometimes saving an object as a string is wise for various reasons including searching. But to answer your question directly, I would go with NSDate predominantly.
I'll stay with Integer 16, because it's the way how it's done in the book "CoreData for iOS".
yes this is an iOS programming question.
I need to calculate the exact time the RMS Titanic sank for its 100 yr anniversary.
It sunk at 15th April 1912 2:20 a.m.
Stupid Question you say. 100th year anniversary is
15th April 1912 2:20 a.m.
+100 years
15th April 2012 2:20 a.m.
But I want an alarm to go off in any timezone in the world exactly and need to handle timezones, and things like British Summer Time being one hour ahead yet it didnt come into effect till 1918.... but as of today 25 mar 2012 London IS in BST so one hour ahead.
Im confused on timezones. We have GMT and UTC. The ships sank at 49° 56' 49" W, 41° 43' 32" N so few hours ahead of London.
-49.94822196927015
41.72713826043066
Whats the correct way to enter a historic date into NSCalendar
and to add 100 years to it exactly and get back the right time
in the users current timezone?
I notice theres Japanese and Islamic calender formats in NSCalendar options. Can iOS device change their dates to these calendars?
And if this was the case. how would I convert from Gregorian to say Islamic?
nice brain puzzler to start your week :)
(There's no code here, as I've never used the iOS API. However, I have some experience of date/time APIs and the oddities they throw up, so I hope you find this answer useful anyway.)
What's the correct way to enter a historic date into NSCalendar
and to add 100 years to it exactly and get back the right time
in the users current timezone?
It really depends on what you mean by "100 years" - it's not like that's a fixed amount of time really.
I would take the UTC instant at which it sank, apply the user's local time zone, and then add a hundred years to that. However, you then need to consider that the result may not actually be a valid local time in that time zone.
For example, suppose you're in a time zone where at the instant of the sinking, it was 1:20am... but in that time zone, 15th April 2012 is when the clocks change - and they skip from 1am to 2am. In that case, 1:20am never occurs... so you could potentially pick 12:20am, the instant before the DST transition, the instant of the DST transition, or 2:20am, depending on what you think is appropriate.
Another possibility to consider is the opposite - suppose it's a DST transition which goes from 2am to 1am... so 1:20am actually occurs twice. What would you want to do in that case? You probably shouldn't make your app celebrate the anniversary twice!
Another option which removes this possibility of ambiguity and discrepancy is to work out what the offset from UTC in the user's time zone was at the exact time of the sinking, then add 100 years to the UTC value (which will never have any DST transitions) and apply the same offset again.
I notice theres Japanese and Islamic calender formats in NSCalendar options. Can iOS device change their dates to these calendars? And if this was the case. how would I convert from Gregorian to say Islamic?
I don't know on that front, I'm afraid.