ts.intersect does not work with xts objects

[joshuaulrich]

Taken from a Stack Overflow question:

The following produces an error

a1 = as.xts(ts(rnorm(20), start=c(1980,1), freq=4))
a2 = as.xts(ts(rnorm(30), start=c(1983,1), freq=4))
a = ts.intersect(a1,a2)

Error in .cbind.ts(list(...), .makeNamesTs(...), dframe = dframe, union = FALSE) : 
  no time series supplied

The documentation says argument should be, "two or more univariate or multivariate time series, or objects which can coerced to time series."

My response:

ts.intersect determines whether the objects is a ts object by looking for the tsp attribute. as.xts.ts removes the tsp attribute, which is why it is not coerced back to a ts object.

This looks like a bug in xts->ts->xts conversion, but I need to take a closer look.

As a work-around, you can manually add the tsp attribute to your xts object (note that this might cause issues with other xts methods, e.g. str.xts) and add a .tsp attribute as well.

set.seed(21)
A1 <- ts(rnorm(20), start=c(1980,1), freq=4)
A2 <- ts(rnorm(30), start=c(1983,1), freq=4)
# convert to xts
a1 <- as.xts(A1)
a2 <- as.xts(A2)
# add tsp attribute
# (so stats:::.cbind.ts will think these are coercible to ts objects)
tsp(a1) <- tsp(A1)
tsp(a2) <- tsp(A2)
# add .tsp attribute
# (needed for as.ts.xts to work)
attr(a1,'.tsp') <- tsp(A1)
attr(a2,'.tsp') <- tsp(A2)

a <- ts.intersect(a1,a2)