Wrong week-ending date using 'to.weekly' function in 'xts' package

Question

I have a really odd issue... I am using the to.weekly and to.period function to convert a daily xts object to weekly data. In most instances, I get the week-ending date as a Friday (day.of.week function will return 5) (e.g. "2010-01-08", "2011-02-11"), but there are a few cases where I get something other than Friday (Saturday/Sunday/Thursday/etc.)

I have tried to.weekly and to.period(x, period = 'weeks') and both return the same problem.

Why is this happening? Is there a work-around for this??

Thanks!!

[EDIT: EXAMPLE BELOW]

test.dates <- as.Date(c("2010-04-27","2010-04-28","2010-04-29","2010-04-30","2010-05-03","2010-05-04","2010-05-05","2010-05-06","2010-05-07","2010-05-10","2010-05-11","2010-05-12","2010-05-13","2010-05-14","2010-05-17","2010-05-18","2010-05-19","2010-05-20","2010-05-21","2010-05-22","2010-05-24","2010-05-25","2010-05-26","2010-05-27","2010-05-28","2010-06-01","2010-06-02","2010-06-03","2010-06-04"))

test.data <- rnorm(length(test.dates),mean=1,sd=2)

test.xts <- xts(x=test.data,order.by=test.dates)

#Function that takes in a vector of zoo/xts objects (e.g. "2010-01-08") and returns the day of the week for each
dayofweek <- function(x) {
 placeholder <- vector("list",length=length(x))
 names(placeholder) <- x

 for(i in 1:length(x)) {placeholder[[i]] <- month.day.year(x[i])}
 placeholder2 <- rep(NA,times=length(x))

 for(i in 1:length(x)) {placeholder2[i] <- day.of.week(placeholder[[i]][[1]],placeholder[[i]][[2]],placeholder[[i]][[3]])}
 return(placeholder2)}

This returns the date(s) that are not Friday: time(to.weekly(test.xts))[dayofweek(time(to.weekly(test.xts))) != 5]

Answer 1

You have 2 problems with your example:

1. Your dayofweek function is a bit cumbersome, and probably incorrect in its results.
2. Your example dates is missing some dates, such as 05-23-2010.

Here is a cleaned-up version of your code:

library(xts)
test.dates <- as.Date(c("2010-04-27","2010-04-28","2010-04-29","2010-04-30","2010-05-03","2010-05-04","2010-05-05","2010-05-06","2010-05-07","2010-05-10","2010-05-11","2010-05-12","2010-05-13","2010-05-14","2010-05-17","2010-05-18","2010-05-19","2010-05-20","2010-05-21","2010-05-22","2010-05-24","2010-05-25","2010-05-26","2010-05-27","2010-05-28","2010-06-01","2010-06-02","2010-06-03","2010-06-04"))
test.data <- rnorm(length(test.dates),mean=1,sd=2)
test.xts <- xts(x=test.data,order.by=test.dates)
test.weekly <- to.weekly(test.xts)

library(lubridate)
test.weekly[wday(test.weekly, label = TRUE, abbr = TRUE) != "Fri"]

The only result of this function is

           test.xts.Open test.xts.High test.xts.Low test.xts.Close
2010-05-22     -1.705749      1.273982    -2.084203      -1.502611

The problem of course, is that this week ends on 05-23-2010, but that date is not present in the time series. Therefore, to.weekly uses the next closest date as the end point, which is 05-22-2010. This is the source of your problem.

Here is a better example, which reveals no issue with the to.weekly function.

library(lubridate); library(xts)   
test.dates <- seq(as.Date("1900-01-01"),as.Date("2011-10-01"),by='days')
test.dates <- test.dates[wday(test.dates)!=1 & wday(test.dates)!=7] #Remove weekends
test.data <- rnorm(length(test.dates),mean=1,sd=2)
test.xts <- xts(x=test.data,order.by=test.dates)
test.weekly <- to.weekly(test.xts)
test.weekly[wday(test.weekly, label = TRUE, abbr = TRUE) != "Fri"]