I wrote a quick function that you can use.

> weekofyear
function(current_date)
{
    t<-as.POSIXlt(current_date)
    t1<-ISOdate(t$year+1900, 1, 1)
    weeks<-difftime(current_date, t1, units="w")
    weeks
}
> weekofyear(as.Date("2007-10-28"))
Time difference of 42.78571 weeks

?weekdays

e.g. weekdays(as.Date('2007-10-31'))

no need for custom functions :)

I am not sure it is you or me. But one of us must have trouble understanding Chinese, English or R language...

oh sorry, I misunderstood his words... but actually only a simple operator %/% is needed...

suppose there're n sample points, then (1:n)%/%7 or rep(1:(n%/%7), each=7) is approximately what he wanted (again, using tapply() & median())