r ifelse date not adding days - r

Question

I need to compute a condition over a column date in R. Atable would be:
PIL_final1<-data.frame( prior_day1_cart=c(4,8),
prior_day1_comp=c('2014-06-03','2014-06-07'),
dia_lim_23_cart=c('201-07-30','201-07-30') )
PIL_final1$prior_day1_comp<-as.Date(PIL_final1$prior_day1_comp, format='%Y-%m-%d')
PIL_final1$dia_lim_23_cart<-as.Date(PIL_final1$dia_lim_23_cart, format='%Y-%m-%d')
So I use ifelse:
PIL_final1$llamar_dia<-ifelse(PIL_final1$prior_day1_cart+6>23,
PIL_final1$dia_lim_23_cart ,
PIL_final1$prior_day1_comp+6)
But I get:
> PIL_final1
prior_day1_cart prior_day1_comp dia_lim_23_cart llamar_dia
1 4 2014-06-03 0201-07-30 16230
2 8 2014-06-07 0201-07-30 16234
And if I do:
> PIL_final1$prior_day1_comp+6
[1] "2014-06-09" "2014-06-13"
I get the right results.
How can I do the ifelse and get the date? thanks.
Also if I try this, I still get a number (although different):
> PIL_final1$llamar_dia<-ifelse(PIL_final1$prior_day1_cart+6>23,
+ PIL_final1$dia_lim_23_cart ,
+ as.Date(PIL_final$prior_day1_comp+6,format="%Y-%m-%d"))
> PIL_final1
prior_day1_cart prior_day1_comp dia_lim_23_cart llamar_dia
1 4 2014-06-03 0201-07-30 16376
2 8 2014-06-07 0201-07-30 16377
Edition:
Also if I do this:
> as.Date(ifelse(PIL_final1$prior_day1_cart+6>23, PIL_final1$dia_lim_23_cart ,
+ PIL_final1$prior_day1_comp+6), format="%Y-%m-%d", origin="1970-01-01")
[1] "2014-06-09" "2014-06-13"
I get the right results, but if I replace the ifelse with the vector result, I get the wrong dates:
> PIL_final1$llamar_dia<-ifelse(PIL_final1$prior_day1_cart+6>23,
+ PIL_final1$dia_lim_23_cart ,
+ PIL_final$prior_day1_comp+6)
> as.Date(PIL_final1$llamar_dia, format="%Y-%m-%d", origin="1970-01-01")
[1] "2014-11-02" "2014-11-03"

Answer 1

from ?ifelse :
The mode of the result may depend on the value of test (see the examples), Sometimes it is better >to use a construction such as
ifelse(test, yes, no) ~~ (tmp <- yes; tmp[!test] <- no[!test]; tmp)
Applying this :
dat$d3 <-
with(dat,{
tmp <- d2+6; tmp[!(x+6>23)] <- d1[!(x+6>23)]; tmp
})
dat
x d1 d2 d3
1 4 2014-06-03 0201-07-30 2014-06-03
2 8 2014-06-07 0201-07-30 2014-06-07
Maybe you should modify this to handle missing values in test.
Note I changed the variables names since yours are really long to type and a real source of errors.
dat <- data.frame( x=c(4,8),
d1=c('2014-06-03','2014-06-07'),
d2=c('201-07-30','201-07-30') )