改进一个 function 来提取年、月和星期几

Question

I'd like to write a wrap function where I can reduce the steps to extract year, month and day of the week, so I don't have to repeartly put down the date variable name. Also if I need extra seasonal variable or some other time object line, I can just add extra line in the function.

The line by line way is

library(lubridate)
yr = year(travel_date)
mth = months(travel_date)
day = weekdays(travel_date))

I can chain them through a dplyr way:

cat <- cat %>% mutate(yr = year(travel_date), 
                      mth = months(travel_date),
                      day = weekdays(travel_date))

The data.table way:

cat[, yr := year(travel_date)][, mnth := months(travel_date)][, dayweek := weekdays(travel_date)]

My wrap function idea is that I only need to put down the dataframe name and the variable name once. But it's not working.

time_spin <- function(x, y) {
  x[, yr := year(y)][,mth := months(y)][,day := weekdays(y)]
}

Any suggestion I can improve this? I can take a wrap function through dplyr or data.table .

dput

cat <- structure(list(id = 1:6, 
                      travel_date = structure(c(16107L, 17266L, 16834L, 17953L, 17953L, 17953L), 
                                              class = c("IDate", "Date"))), 
                 row.names = c(NA, -6L), 
                 class = c("data.table", "data.frame"))

Answer 1

Using dplyr you can write the function as:

time_spin <- function(x, y) {
  x %>% mutate(yr = year(.data[[y]]), 
                 mth = months(.data[[y]]),
                 day = weekdays(.data[[y]]))
}

time_spin(cat, 'travel_date')

#   id travel_date   yr      mth       day
#1:  1  2014-02-06 2014 February  Thursday
#2:  2  2017-04-10 2017    April    Monday
#3:  3  2016-02-03 2016 February Wednesday
#4:  4  2019-02-26 2019 February   Tuesday
#5:  5  2019-02-26 2019 February   Tuesday
#6:  6  2019-02-26 2019 February   Tuesday