使用 lubridate 和管道将日期转换为年月日

Question

I want to turn a date in "day/month/year" to "year/month/day" using lubridate and pipes but I'm having problems. Also, the first row of 'public_release' has day and month switched. The first three rows of both bwg_release and public_release should reflect a January 6 date. Eg

    datasets %>%
+   select(bwg_release, public_release)
  bwg_release public_release
1  06/01/2016     01/06/2018
2  06/01/2016     06/01/2018
3  06/01/2016     06/01/2018
4  17/04/2016     17/04/2016

Current output:

datasets %>% 
+   mutate(bwg_release_new_date = format(dmy(bwg_release),"%Y/%m/%d")) %>%
+   mutate(public_release_new_date = format(dmy(public_release),"%Y/%m/%d")) %>%
+   select(bwg_release_new_date,public_release_new_date)
  bwg_release_new_date public_release_new_date
1           2016/01/06              2018/06/01
2           2016/01/06              2018/01/06
3           2016/01/06              2018/01/06
4           2016/04/17              2016/04/17     

Desired output:

bwg_release_new_date public_release_new_date
1           2016/01/06              2018/01/06
2           2016/01/06              2018/01/06
3           2016/01/06              2018/01/06
4           2016/04/17              2016/04/17

Answer 1

I believe the recommendation is to use the base format function. Just read the date with dmy() and then define your format.

library(dplyr)
library(lubridate)
datasets %>% 
  mutate(output = format(dmy(bwg_release), format = "%Y/%m/%d"))
  bwg_release     output
1  06/01/2016 2016/01/06
2  06/01/2016 2016/01/06
3  06/01/2016 2016/01/06
4  17/04/2016 2016/04/17

Answer 2

For the sake of completeness Using regular expression although @IanCampbell (+1) solution is the way to go.

datasets %>%
 mutate(bwg_release=ymd(gsub("(\\d+)/(\\d+)/(\\d+)","\\3/\\2/\\1",  bwg_release)))
# A tibble: 4 x 1
  bwg_release
  <date>     
1 2016-01-06 
2 2016-01-06 
3 2016-01-06 
4 2016-04-17