How to align Dates from different columns in R or Excel - r

Question

I have a dataset of variables with each variable having a different date span. They are presented as in the following example (taken the first two cases out of 500):
DatesV1 DatesV2
29/12/1995 19/07/2001
02/01/1996 20/07/2001
03/01/1996 23/07/2001
04/01/1996 24/07/2001
05/01/1996 25/07/2001
08/01/1996 26/07/2001
09/01/1996 27/07/2001
10/01/1996 30/07/2001
11/01/1996 31/07/2001
12/01/1996 01/08/2001
15/01/1996 02/08/2001
16/01/1996 03/08/2001
17/01/1996 06/08/2001
What I want to happen is for the dates in DatesV2 to align with the dates in DatesV1. This means that DatesV2 will start with a few NA until the row that the dates align. Like this:
DatesV1 DatesV2
... ...
17/07/2001 NA
18/07/2001 NA
19/07/2001 19/07/2001
20/07/2001 20/07/2001
... ...
In the Example Set, I have the example of exactly what I am trying to do. I can't find of a fast computational way to do it in either R or Excel for the 500 variables that I have.
Example Set
I have tried something like this:
nhat<-which(Example$DatesV2[1]==Example$DatesV1)
nend<-which(Example$DatesV1[length(Example$DatesV1)-1]==Example$DatesV2)
Example$Apotelesma<- c(rep(NA,nhat-1),Example$DatesV2[1:nend],NA)
Which is an initial test that works for two dates. Only thing is that dates appear as numbers.

Answer 1

Here's a possible solution using some re-shaping. I'm using a simple example:
df = data.frame(DatesV1 = c("24/07/2001","25/07/2001","26/07/2001"),
DatesV2 = c("25/07/2001","26/07/2001","27/07/2001"),
DatesV3 = c("26/07/2001","27/07/2001","28/07/2001"),
stringsAsFactors = F)
library(tidyverse)
library(lubridate)
# update to date columns (only if needed)
df = df %>% mutate_all(dmy)
df %>%
gather() %>% # reshape dataset
mutate(id = value) %>% # use date values as row ids
spread(key, value) %>% # reshape again
select(-id) # remove ids
# DatesV1 DatesV2 DatesV3
# 1 2001-07-24 <NA> <NA>
# 2 2001-07-25 2001-07-25 <NA>
# 3 2001-07-26 2001-07-26 2001-07-26
# 4 <NA> 2001-07-27 2001-07-27
# 5 <NA> <NA> 2001-07-28

Answer 2

This an ugly/messy method if you like, but it gets the job done. Any faster and tidier way would be better.
n<-nrow(DataAlignment)
Newdata<-matrix(0,5148,ncol(DataAlignment))
loops<-ncol(DataAlignment)-1
for(i in 1:loops){
nhat<-which(DataAlignment[1,i+1]==DataAlignment[,1]) #finds the position of the first date in column 2 according to the first column
nend<-which(DataAlignment[n,1]==DataAlignment[,i+1]) #finds the position of last date in col 2 according to the first column
if(nhat==1 | nend != 5148){ #takes into account when they start at the same time but end in different dates
Newdata[,i+1]<-c(DataAlignment[c(1:nend),i+1],rep(NA,n-nend))
}
else{if(nhat==1| nend==5148){Newdata[,i+1]<-c(DataAlignment[,i+1])} #this takes account when they start and end at the same time
else{if(nhat!=1){
Newdata[,i+1]<-c(rep(NA,nhat-1),DataAlignment[c(1:nend),i+1])}}} #creates the new data
}