create and export as.nanotime

[jangorecki]

It might be sometimes useful to operate on arbitrary type objects. For example if fread's colClasses arg will be supplied with "nanotime" type, then we could flexibly dispatch to arbitrary class by using

convert = function(x, type) {
  as.call(c(as.name(paste("as", type, sep=".")), as.name("x")))
}
convert(x, type="nanotime")
#as.nanotime(x)

as.* functions are kind of standard in R.

[lsilvest]

If I understand correctly, the end goal would be something like this, with a conversion from character to nanotime?

fread("a_file_that_contains_nanotimes.csv", colClasses="nanotime")

[jangorecki]

fread was just an example of a function that could utilise as.nanotime.
source data type doesn't matter as long as there are methods to dispatch, as.nanotime.character, and most importantly as.nanotime.deafult.

[eddelbuettel]

For reasons I no longer remember our constructor was always just nanotime() so we could possibly simply alias and register. We do register S3 methods for the other direction:

S3method(as.Date,nanotime)
S3method(as.POSIXct,nanotime)
S3method(as.POSIXlt,nanotime)
S3method(as.data.frame,nanotime)
S3method(as.integer64,nanotime)

[lsilvest]

The reason I was asking for the end goal is that I'm not sure if there's a consensus that as.to_type.from_type be provided. The generic coercion function for an S4 class is as (by the way not yet implemented either for nanotime). So I don't think the convert function can expect full generality as is, especially for S4 classes. I'll try to dig into this a bit deeper...

[jangorecki]

@lsilvest regarding fread and as.* methods I mentioned, related point Rdatatable/data.table@1de0399#diff-8312ad0561ef661716b48d09478362f3R11

colClasses now supports 'complex', 'raw', 'Date', 'POSIXct', and user-defined classes (so long as an as. method exists)

[jangorecki]

example where as.nanotime could be helpful: Rdatatable/data.table#4101

[eddelbuettel]

I am sorry but can you be more specific? I read the #4101 link but it "didn't click".

[jangorecki]

Provided wrapper function calls as.numeric / as.integer / etc based on values in colClasses by calling do.call(paste0("as.", colClasses[[i]]), ...), thus if user wants a column to be nanotime then colClasses(time="nanotime", ...) allows to automatically convert a column to nanotime.

[lsilvest]

Yes, I agree we should provide the as.nanotime methods. I think Dirk's suggestion to alias should work well. Furthermore, I think we should in parallel provide the as type coercion via setAs functions. This would provide generality in type conversion for nanotime whichever way a users chooses.