Using ifelse statement when concatenating elements of a date variable - r

Question

I am attempting to use two ifelse statements to create a new date variable that makes a series of assumptions to fill in the gaps of an existing date variable. Here is an example of what I mean:
id EffectiveDate EffectiveYear ED_NA EY_NA NewEffectiveDate
1 a 1972-10-05 1972 FALSE FALSE 1972-10-05
2 a <NA> 1985 TRUE FALSE 1985-01-01
3 a 1988-11-12 1988 FALSE FALSE 1988-11-12
4 b 2011-09-05 2011 FALSE FALSE 2011-09-05
5 b <NA> NA TRUE TRUE 2011-09-05
6 b <NA> 2012 TRUE FALSE 2012-01-01
7 c 2012-11-11 2012 FALSE FALSE 2012-11-11
8 c 2013-05-15 2013 FALSE FALSE 2013-05-15
quick code for id:EY_NA =
id <- c("a","a","a","b","b","b","c","c")
EffectiveDate <- c("1972-10-05",NA,"1988-11-12","2011-09-05",NA,NA,"2012-11-11","2013-05-15")
EffectiveYear <- c(1972,1985,1988,2011,NA,2012,2012,2013)
tdat <- data.frame(id, EffectiveDate, EffectiveYear)
tdat$ED_NA <- is.na(tdat$EffectiveDate)
tdat$EY_NA <- is.na(tdat$EffectiveYear)
What I'm trying to create in this example is the "NewEffectiveDate" variable. In plain English, what I want is, where EffectiveDate data are missing BUT EffectiveYear data are not missing, assume NewEffectiveDate is equal to January 1 of the EffectiveYear. If EffectiveDate AND EffectiveYear data are missing, assume the prior observation's EffectiveDate. Last, of course, if EffectiveDate data are not missing, select EffectiveDate.
Here is the latest code I used to attempt to solve the problem:
tdat %>% mutate(NewEffectiveDate = ifelse(ED_NA == 1 & EY_NA == 0,
as.Date(paste(EffectiveYear, 1, 1, sep="-")),
ifelse(ED_NA == 1 & EY_NA == 1),
as.Date(lag(EffectiveDate)),
EffectiveDate
))
When I try this particular code, I get an error message that reads: Error: unused arguments (as.Date(c(NA, 1, NA, 2, 3, NA, NA, 4)), c(1, NA, 2, 3, NA, NA, 4, 5))
I searched for similar questions with queries like "ifelse concatenate date" and some variations thereof, but haven't found anything that seems to apply to this particular problem.
I am very new to R (and CLIs, for that matter), so I apologize in advance if I'm overlooking a perfectly obvious solution. The transition from Excel to R has been interesting, but often painful when it comes to doing what seem like relatively straightforward tasks (though the dplyr package has been tremendously helpful).

Answer 1

id <- c("a","a","a","b","b","b","c","c")
EffectiveDate <- c("1972-10-05",NA,"1988-11-12","2011-09-05",NA,NA,"2012-11-11","2013-05-15")
EffectiveYear <- c(1972,1985,1988,2011,NA,2012,2012,2013)
tdat <- data.frame(id, EffectiveDate, EffectiveYear,
stringsAsFactors=FALSE)
library(zoo)
tdat %>%
mutate(NewEffectiveDate = ifelse(!is.na(EffectiveDate),
EffectiveDate,
ifelse(is.na(EffectiveDate) & !is.na(EffectiveYear),
paste0(EffectiveYear, "-01-01"),
NA)),
NewEffecitveDate = na.locf(NewEffectiveDate))
This should give you what you need. I recommend using na.locf (last one carried forward) from the zoo package rather than trying to deal with the previous date issue.

Answer 2

You can do
tdat$EffectiveDate <- as.Date(tdat$EffectiveDate)
tdat %>% mutate(NewEffectiveDate = as.Date(
ifelse(!is.na(EffectiveDate), EffectiveDate,
ifelse(!is.na(EffectiveYear), as.Date(paste(EffectiveYear, 1, 1, sep="-")),
lag(EffectiveDate)))
)) -> res
res
# id EffectiveDate EffectiveYear NewEffectiveDate
# 1 a 1972-10-05 1972 1972-10-05
# 2 a <NA> 1985 1985-01-01
# 3 a 1988-11-12 1988 1988-11-12
# 4 b 2011-09-05 2011 2011-09-05
# 5 b <NA> NA 2011-09-05
# 6 b <NA> 2012 2012-01-01
# 7 c 2012-11-11 2012 2012-11-11
# 8 c 2013-05-15 2013 2013-05-15

Answer 3

There appears to be a problem with your ifelse block you closed the bracket for the second block early and didn't give a yes or no argument and you gave an extra argument to the first ifelse block.
This should work:
tdat %>% mutate(NewEffectiveDate = ifelse(ED_NA == 1 & EY_NA == 0,
as.Date(paste(EffectiveYear, 1, 1, sep="-")),
ifelse(ED_NA == 1 & EY_NA == 1,
as.Date(lag(EffectiveDate))),
EffectiveDate))