Ulrich Kohler
> I have two questions on elapsed dates:
>
> The mdy()-function returns the elapsed date, counting from
> January 1 1960.
>
> . display mdy(1,1,1960)
> 0
>
> . display mdy(1,2,1960)
> 1
>
> . display mdy(12,24,100)
> -678993
>
> However
>
> . display mdy(12,24,99)
>
> returns a missing value, as does
>
> . display mdy(12,24,00)
> .
>
> What is the reason that mdy() returns missing with 2 digit
> years ( to prevent
> the user from using two digit years?)
Dates in Stata start at 1 January 100. That once put paid
to a project I wanted to do on modelling of
daily temperature measurements and olive prices under
Julius Caesar, but Stata Corp were unyielding: they have
no interest in changing this limit.
As you say, elsewhere you can indicate separately a two-digit
year and a century, so that "52" and "19" can be coupled
together. It gets a bit awkward to imagine "52" and "0"
being coupled together to indicate the year 52.
> The second question is just for curiosity. Is there any
> reason for using the
> January 1 1960 as zero? It seems quite common in software
> programs, but why?
Interesting little historical question. I guess some
data base or spreadsheet program started this as a
convention, and lots followed. Perhaps there was a
presumption that nobody much wanted to deal with
data before that. (Depending upon implementation,
perhaps you could have dates before as negative numbers,
as in Stata.)
Also, suppose you used 1 January 1 as a basis. Then
most daily dates which would be used would be represented
by integers ~ 700,000 and that could tie up lots of bytes which,
a few years ago, was of course a much bigger deal.
As far as Stata is concerned, a more natural base
would be 21 January 1952 (see [U] p.302), but that
would be rather too idiosyncratic to fit the bill.
Nick
[email protected]
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
