nanotime(NA) should not return a nanotime

[lsilvest]

> nanotime(NA)
[1] "1677-09-21T00:12:43.145224192+00:00"

[lsilvest]

It's actually most likely a formatting issue, but is.na also needs to be defined for nanotime:

> is.na(nanotime(NA))
[1] TRUE
attr(,".S3Class")
[1] "integer64"
attr(,"class")
[1] "nanotime"
attr(,"class")attr(,"package")
[1] "nanotime"

[eddelbuettel]

Ditto for other non-finite values it seems:

R> nanotime(c(-Inf, NA, NaN, Inf))
[1] "1677-09-21T00:12:43.145224192+00:00" "1677-09-21T00:12:43.145224192+00:00" 
[3] "1677-09-21T00:12:43.145224192+00:00" "1677-09-21T00:12:43.145224192+00:00"
Warning message:
In as.integer64.double(x, keep.names = TRUE) :
  NAs produced by integer64 overflow
R> 

The warning comes from the Inf values.

[lsilvest]

Yes, you are right, we need to take care of these other cases too. I'll get a patch together for that.

[eddelbuettel]

Now fixed:

R> nanotime(c(-Inf, NA, NaN, Inf))
[1] NA NA NA NA
Warning message:
In as.integer64.double(x, keep.names = TRUE) :
  NAs produced by integer64 overflow
R>