Doesn't detect dates

Question

I am doing some operations with the functions isoweek and if_else but some dates do not perform the operation obtaining NA.

db = read_dta("C:/Users/crist/Downloads/db.dta")


db$date1 = substr(db$date1,1,10)
db$date1 = as.Date(db$date1)
db$date2 = substr(db$date2,1,10)
db$date2 = as.Date(db$date2)

t = db %>% 
  drop_na(value) %>% 
  mutate(Weekday = weekdays(date2),
         date2 = replace(date2, Weekday %in% c("Saturday", "Sunday"), NA),
         num_week = isoweek(date2),
         dummy_sas = if_else(date2 >= ymd("2020-05-18"),1,0)) %>% 
  fill(date2)  %>% 
  select(-Weekday) %>% 
  mutate(date2 = if_else(date2 < as.Date("2020-05-18") & SAS == 1, date1, date2))

I need for example all the week numbers but i get NA's

> summary(isoweek(t$date2))
   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
   1.00   22.00   31.00   28.15   39.00   47.00 
> summary(t$num_week)
   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
   1.00   21.00   31.00   28.12   39.00   47.00     109 

Data here

Answer 1

I haven't downloaded the data but I think order of operation seems to be an issue here. Try this :

library(dplyr)
library(lubridate)

db %>% 
  drop_na(value) %>% 
  mutate(Weekday = weekdays(date2),
         num_week = isoweek(date2),
         date2 = replace(date2, Weekday %in% c("Saturday", "Sunday"), NA),
         dummy_sas = as.integer(date2 >= ymd("2020-05-18"))) %>%
  fill(date2)  %>% 
  select(-Weekday) %>% 
  mutate(date2 = if_else(date2 < as.Date("2020-05-18") & SAS == 1, date1, date2))

---

In your code you replace weekends with NA :

date2 = replace(date2, Weekday %in% c("Saturday", "Sunday"), NA)

and then take isoweek on date2.

num_week = isoweek(date2)

So weekends would give NA in num_week.

You replace date2 again in last line

mutate(date2 = if_else(date2 < as.Date("2020-05-18") & SAS == 1, date1, date2))

so you don't get any NAs in t$date2.