Generate alternate saturday dates list within range in R

Question

library(tidyr)
d = seq(as.Date("2018-01-01"),Sys.Date()+365,by='day')
saturdayList = d[weekdays(d)=='Saturday']

altSaturdayList = data.frame(lapply(split(saturdayList, format(saturdayList, "%Y-%m")), function(x)
  na.omit(x[c(2,4)])))

altSaturdayList = subset(gather(altSaturdayList),select = c(value))
HolidayList = rbind(HolidayList, setNames(altSaturdayList, "Holidays"))
HolidayList = rbind(HolidayList, setNames(as.data.frame(as.Date("2018-11-06")), "Holidays"))
HolidayList = arrange(unique(HolidayList),Holidays)

rm(d)
rm(saturdayList)
rm(altSaturdayList)

The above code generates a list of dates which contain 2nd and 4th saturday of every month within that date range.

As you can tell I have generated and deleted intermediate variables like d, saturdayList and altSaturdayList. I was wondering how I can make this code better and efficient?

I learned about %>% operator but I don't know how I can do it with no input.

Any tips to improve the code is welcomed. Thanks

Answer 1

A pipe (%>%) solution is the following. I also use the dplyr-package, which you are probably already using as you have the arrange-function in your code. First I create a tibble with the days as input and then we can pipe ahead ;)

library(tidyr)
library(dplyr)

dates <- tibble(days = seq(as.Date("2018-01-01"), Sys.Date() + 365, by = 'day'))

HolidayList <- dates %>% 
  # Only keep Saturdays
  filter(weekdays(days) == 'Saturday') %>% 
  # Create a group variable (year_month) within we want to select the 2nd and 4th value
  group_by(year_month = format(days, "%Y-%m")) %>% 
  # Only keep the 2nd and 4th value in each year_month
  slice(c(2, 4)) %>% 
  # Get rid of the grouping 
  ungroup() %>% 
  # Only select the days column and rename it to Holidays
  select(Holidays = days) %>% 
  # Add the extra holiday 2018-11-06
  add_row(Holidays = as.Date("2018-11-06")) %>% 
  # similar to unique.data.frame(), but considerably faster:
  distinct() %>% 
  # Arrange ascending by Holidays
  arrange(Holidays)

HolidayList
# A tibble: 54 x 1
#      Holidays  
#        <date>    
#  1 2018-01-13
#  2 2018-01-27
#  3 2018-02-10
#  4 2018-02-24
#  5 2018-03-10
#  6 2018-03-24
#  7 2018-04-14
#  8 2018-04-28
#  9 2018-05-12
# 10 2018-05-26
# ... with 44 more rows