missing `date` function

[joethorley]

lubridate provides day, hour and year functions to get and set the day, hour and year components of a date time. A logical addition would be a date function to get and set the date component. It would circumvent the use of as.Date(,tz) which can cause problems if tz is set wrong. It would conflict with the base version of date which is poorly named and effectively replaced by lubridate::today. It could be called d8 to avoid any conflicts.

Please let me know if I am missing something.

I am willing to implement.

[vspinu]

Closely related to #355.

I would call it date for the sake of consistency. Users of lubridate are very unlikely to use base::date anyways.

@hadley, @huftis, what do you think?

[hadley]

Hmmm, I'm not a big fan of overriding base R functions (although I agree it wouldn't make much difference here). Why not call it make_date() to be consistent with make_datetime()?

[vspinu]

Why not call it make_date() to be consistent with make_datetime()?

make_datetime is a constructor. The proposal is to add a "getter" and "setter" for date. That is, to extend the existent family of our accessors -year, day, month etc.

[joethorley]

get_date and set_date could work?

[hadley]

Ooooh, I totally misread that

[joethorley]

Unless someone tells me otherwise I'm going to do a pull request early next week for get_date and set_date

[vspinu]

get_date and set_date is pretty bad IMO. set_date conflicts with R's foo<- convention and makes our accessors heterogeneous, which I don't like at all.

Overwriting date is a small price to pay for clarity and consistency of date(x) and date(x) <- d. Let's just go with overwrite for now and change it in a later stage if we happen to have a better idea.

[joethorley]

Agreed - I'll implement early next week

Sent from my iPhone

On Jan 22, 2016, at 22:02, Vitalie Spinu notifications@github.com wrote:

get_date and set_date is pretty bad IMO. It conflicts with R's foo<-
convention and makes our accessors heterogeneous, which I don't like at
all.

Overwriting date is a small price to pay for clarity and consistency of
date(x) and date(x) <- d. Let's just go with overwrite for now and change
it in a later stage if we happen to have a better idea.

—
Reply to this email directly or view it on GitHub
#373 (comment).