Question

I’m getting unexpected output from the all.equal method in R, specifically the implementation for POSIXct, all.equal.POSIXct.

t <- Sys.time()
isTRUE(all.equal(t, t+1))

returns TRUE, and

isTRUE(all.equal(t, t+1, scale = 1))

returns FALSE.

However, if you look at the definition of all.equal.POSIXct, you can see that the scale parameter has a default of 1:

> all.equal.POSIXct
function (target, current, ..., scale = 1) 
{
    check_tzones(target, current)
    NextMethod("all.equal")
}
<bytecode: 0x22eac90>
<environment: namespace:base>

You get the same results if you explicitly call all.equal.POSIXct instead of all.equal.

Why isn’t the default parameter scale = 1 being picked up in the first call to all.equal.POSIXct? Am I doing something wrong, or have I fundamentally misunderstood something, or is this a bug?

Thanks in advance for any help.

Answer 1

I’m going out on a slight limb here, but I think you have discovered a bug.

Here is my suggested fix:

all.equal.POSIXct <- function (target, current, ..., scale = 1) {
  check_tzones(target, current)
  NextMethod("all.equal", scale=scale, ...)
}

Then the function gives the correct results:

all.equal(t, t+1)
[1] "Mean scaled difference: 1"

all.equal(t, t+1, scale=10)
[1] "Mean scaled difference: 0.1"

---

This is why the existing code doesn’t work:

The definition for all.equal is:

all.equal <- function (target, current, ...) UseMethod("all.equal")

• Notice that there are three arguments: target, current and ....

• Thus, whenever you use NextMethod these three arguments will be passed to the next method.

• However, in the case of all.equal.POSIXct there is an additional argument scale=, but this doesn’t get passed on either implicitly or explicitly.