R function that returns value based on Date?

Question

I have 2 datasets, each with 2 columns. In the both sets, the columns are Date and Value. Dataset 1 are actual recorded values where the last Date is 06-10. Dataset 2 are predicted values for all the Dates in Dataset 1, with additional rows for future dates (6-11, 6-12, 6-13, etc.).

For example,

d1_date <- c('2021-06-08', '2021-06-09', '2021-06-10')
d1_value <- c(1, 2, 3)
d1 <- data.frame(d1_date, d1_value)

d2_date <- c('2021-06-08', '2021-06-09', '2021-06-10', '2021-06-11', '2021-06-12', '2021-06-13')
d2_value <- c(3, 2, 5, 4, 5, 6)
d2 <- data.frame(d2_date, d2_value)

Now I have a list of dates ranging from '2021-06-08' to '2021-06-13' and I want to put d1 values into the dates 06-08 to 06-10, and the values from d2 into 06-11 to 06-13, so that the corresponding values would be 1, 2, 3, 4, 5, 6 for the new data frame.

I have a function below:

fb_fb3 <- function(date) {
  
  date <- as.Date(date)
  
  ifelse(date <= max(d1$d1_date), return(d1$d1_value[d1$d1_date %in% date]), return(d2$d2_value[d2$d2_date %in% date]))

}

This somewhat works, as when I put in the list of the dates from 06-08 to 06-13 in the function, it gives all the values from whichever dataframe I mention first in the ifelse.

When I put in d1 first, I can't bind it because it only returns 3 values, not 6. When I put in d2, I can bind it but the values will be 3,2,5,4,5,6.

I'm unable to combine values from d1 and d2 to get 1,2,3,4,5,6. Can someone please give some advice? Thanks!

Answer 1

We could use if/else and then Vectorize the function

fb_fb3 <- function(date) {
  if(!inherits(date, 'Date'))
    date <- as.Date(date)
  if(!inherits(d1$d1_date, "Date"))
    d1$d1_date <- as.Date(d1$d1_date)
  if(!inherits(d2$d2_date, "Date"))
    d2$d2_date <- as.Date(d2$d2_date)
    
    
  if(date <= max(d1$d1_date)) 
    with(d1, d1_value[d1_date %in% date])
  else 
     with(d2, d2_value[d2_date %in% date])
 }

-input dates

dates <- seq(as.Date( '2021-06-08'), as.Date('2021-06-13'), by = '1 day')  

-testing

Vectorize(fb_fb3)(dates)
#[1] 1 2 3 4 5 6

Answer 2

How about using a join and coalesce ?

library(dplyr)

d1 %>%
  full_join(d2, by = c('d1_date' = 'd2_date')) %>%
  transmute(date = d1_date, 
            value = coalesce(d1_value, d2_value))

#        date value
#1 2021-06-08     1
#2 2021-06-09     2
#3 2021-06-10     3
#4 2021-06-11     4
#5 2021-06-12     5
#6 2021-06-13     6