> On Fri, Sep 2, 2011 at 9:59 AM, Marc Schwartz <marc_schwa...@me.com> wrote:
>
>> On Sep 2, 2011, at 10:35 AM, Ben qant wrote:
>>
>>> Hello,
>>>
>>> I'm attempting to return the date (in form '%Y-%m-%d') of the Monday
>>> previous to the current date. For example: since it is 2011-09-02 today,
>> I
>>> would expect 2011-08-29 to be the return value.
>>>
>>> I found the following in:
>>> http://www.mail-archive.com/r-help@r-project.org/msg144184.html
>>>
>>> Start quote from link:
>>> prevmonday <- function(x) 7 * floor(as.numeric(x-1+4) / 7) + as.Date(1-4)
>>>
>>> For example,
>>>
>>>> prevmonday(Sys.Date())
>>> [1] "2011-08-15"
>>>> prevmonday(prevmonday(Sys.Date()))
>>> [1] "2011-08-15"
>>>
>>> End quote from link.
>>>
>>> But when I do it I get:
>>>> prevmonday <- function(x) 7 * floor(as.numeric(x-1+4) / 7) +
>> as.Date(1-4)
>>>> prevmonday(Sys.Date())
>>> Error in as.Date.numeric(1 - 4) : 'origin' must be supplied
>>>
>>> I've tried setting the 'origin' argument in as.Date() in different ways,
>> but
>>> it returns inaccurate results.
>>>
>>> Thanks,
>>>
>>> Ben
>>
>>
>> If memory serves, this is because Gabor used the version of as.Date() from
>> his 'zoo' package in that post, which does not require an origin to be
>> specified, whereas the default as.Date() function in R's base package does:
>>
>> prevmonday <- function(x) 7 * floor(as.numeric(x-1+4) / 7) + as.Date(1-4)
>>
>>> prevmonday(Sys.Date())
>> Error in as.Date.numeric(1 - 4) : 'origin' must be supplied
>>
>>> require(zoo)
>> Loading required package: zoo
>>
>> Attaching package: 'zoo'
>>
>> The following object(s) are masked from 'package:base':
>>
>> as.Date
>>
>>> prevmonday(Sys.Date())
>> [1] "2011-08-29"
>>
>>
>> # Remove 'zoo' to use the base function
>> detach(package:zoo)
>>
>>> prevmonday(Sys.Date())
>> Error in as.Date.numeric(1 - 4) : 'origin' must be supplied
>>
>>
>> # Fix the function to use base::as.Date()
>> prevmonday <- function(x) 7 * floor(as.numeric(x-1+4) / 7) + as.Date(1-4,
>> origin = "1970-01-01")
>>
>>> prevmonday(Sys.Date())
>> [1] "2011-08-29"
>>
>>
>> See ?as.Date
>>
>> HTH,
>>
>> Marc Schwartz
On Sep 2, 2011, at 11:02 AM, Ben qant wrote:
> Oh OK, missed that.
>
> Here is a solution using base: (already posted)
>
> I didn't sort out the issue in my email below but here is a (not very
> R'ish) solution:
>
>> pm = function(x) {
> + for(i in 1:7){
> + if(format(as.Date(Sys.Date()-
> i),'%w') == 1){
> + d = Sys.Date() - i;
> + }
> + }
> + d
> + }
>> pm(Sys.Date())
> [1] "2011-08-29"
It occurs to me that another solution would be:
> as.Date(cut(Sys.Date(), "weeks"))
[1] "2011-08-29"
This uses cut.Date() to create a vector that contains the beginning of weekly
intervals for each date value, with the week beginning on a Monday by default.
Note that cut.Date() returns a factor, not a Date class object, hence the
additional coercion of the result.
See ?cut.Date
So if you want that in a function wrapper, you could do:
prev.monday <- function(x) as.Date(cut(x, "weeks"))
> prev.monday(Sys.Date())
[1] "2011-08-29"
HTH,
Marc
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
