Fix Datetimediff () bug with "months" and leap day

[denny]

There is a bug in DateTimeDiff(a, b, "months") whereby if a is a leap day DateTimeDiff() returns a value by less than 1 month.

Please see attached workflow where I demonstrate this bug.

Alteryx95DateTimeDiff bug and workaround.yxmd

[RachelW]

We'll look into this. What version are you using? 

[denny]

@RachelW I'm currently using 11.5 but this bug was first discovered on 9.5, which is when I filed a bug report on 2016-06-11.

Somehow the response replied with this is a feature and not a bug.

 I haven't run the attached workflow against 2018.1 release yet.

The original support case number is 00114317.

 The current workaround I use is to convert the dates involved to their respective 1st of the month, prior to running the DateTimeDiff() function.

[RachelW]

@denny Thanks! We're looking into it. 

[RachelW]

@denny Thank you for discussing this issue with me further. We will update our Help documentation to be clear as to how we are making the calculation and include how that impacts the calculation in the case of starting a time period with a leap day. 

[RachelW]

Our Help documentation has been updated to make the details of this function more clear. Thanks for submitting it to the ideas forum. 

[denny]

@RachelW:

Thank you. I see the documentation update at https://help.alteryx.com/current/Reference/Functions.htm

Just wanted to mention I'm still requesting Alteryx to update the circled example to 12 instead of 11, through say an optional parameter. I understand Alteryx conforms to the Lotus 1-2-3 standards that DateTimeDiff() is currently using, and should be kept this way to avoid breaking existing Alteryx workflows, but Alteryx should also support a more common and standard interpretation than the 1980s Lotus 1-2-3 standard.