[R-sig-Epi] FW: transforming number from cal.yr() back to Date

[BXC (Bendix Carstensen)]

I got this problem presented, and I think it is of general interest for the
list. I copy it by agreement with Tina.

Best,
Bendix

-----Original Message-----
From: BXC (Bendix Carstensen)
Sent: Monday, February 13, 2006 6:54 PM
To: 'Zagar Tina'
Subject: RE: transforming number from cal.yr() back to date


> -----Original Message-----
> From: Zagar Tina [mailto:TZagar at onko-i.si]
> Sent: Monday, February 13, 2006 3:42 PM
> To: BXC (Bendix Carstensen)
> Subject: transforming number from cal.yr() back to date
>
>
> Dear Bendix,
>
> I am working with function Lexis() included in Epi package so I have
> to transform all dates into numbers using function cal.yr(). I am
> having problems with transforming these numbers back to dates. I have
> tried with transforming cal-yr()'s number to R date first but
> obviously I am doing something terribly wrong. Can you help me? I am
> sending you my code (which does not work):
>
> # the original date is 07.04.1994
> a = cal.yr("07.04.1994" , format="%d.%m.%Y")
> a # result is 1994.263
> # 'a1' is 'a' transformed into R date (number of days)
> a1 = round(as.numeric(as.Date( as.character(paste("01.01.",
> round(a))) , format="%d.%m.%Y"))+ 365.25 * (a - round(a))) # 'a2' is R
> date transformed from original date 07.04.1994 a2 =
> as.numeric(as.Date( as.character("07.04.1994") ,
> format="%d.%m.%Y")) # a1 and a2 are the same a1 - a2 # result is 0 #
> but converting a1 and a2 back to ordinary dates does not work
> as.date(a1) # result is 6Apr84
> as.date(a2) # result is 6Apr84
>
> I hope my question is not too stupid... Mybe only last step is false
> "as.date(a1)"? Thank you for your answer.
>
> Best regards,
> Tina Zagar.
>
> <><><><><><><><><><><><><><>
> Tina Žagar
> Institute of Oncology Ljubljana
> Epidemiology and Cancer Registries
> Zaloška cesta 2
> SI-1000 Ljubljana
> Slovenia
> E-mail: tzagar at onko-i.si
> Tel.: +386 1 5879 562
> <><><><><><><><><><><><><><>

Dear Tina,

I was a bit surprised myself, because if you do:

> methods(as.Date)
[1] as.Date.character as.Date.date      as.Date.dates     as.Date.default 
[5] as.Date.factor    as.Date.POSIXct   as.Date.POSIXlt 

there is no as.Date.numeric. But a date is merely the number of days since
1 Jan 1970, equipped with the class "Date", so below is two ways of doing it:

> # the original date is 07.04.1994
> a = cal.yr("07.04.1994" , format="%d.%m.%Y")
> a # result is 1994.263
[1] 1994.263
> b <- structure((a-1970)*365.25,class="Date")
> b
[1] "1994-04-07"
> bb <- (a-1970)*365.25
> class(bb) <- "Date"
> bb
[1] "1994-04-07"

Best,
Bendix
----------------------
Bendix Carstensen
Senior Statistician
Steno Diabetes Center
Niels Steensens Vej 2
DK-2820 Gentofte
Denmark
tel: +45 44 43 87 38
mob: +45 30 75 87 38
fax: +45 44 43 07 06
bxc at steno.dk
www.biostat.ku.dk/~bxc