Fill blank rows depends on previous/next non-empty value

Question

I have a subscription data frame looks like the following. There are about 1 million unique IDs. The table lists subscription status. When user began subscription service, status field is indicated by 'Sub' and when user unsubscribed, it's indicated by 'Usub'.

dat <- data.frame(ID=c(rep("A",12),(rep("B",12))), Year="2014", Month=rep(seq(1:12),2), Status=NA)
dat$Status[4] <- "Sub"
dat$Status[8] <- "Usub"
dat$Status[17] <- "Usub"
dat$Status[21] <- "Sub"

ID Year Month Status
A  2014   1    
A  2014   2    
A  2014   3    
A  2014   4    Sub
A  2014   5    
A  2014   6    
A  2014   7    
A  2014   8    Usub 
A  2014   9    
A  2014  10    
A  2014  11    
A  2014  12    
B  2014   1    
B  2014   2    
B  2014   3    
B  2014   4    
B  2014   5    Usub    
B  2014   6    
B  2014   7    
B  2014   8    
B  2014   9    Sub
B  2014  10    
B  2014  11    
B  2014  12    
C  2014   1     .
.    .    .     .
.    .    .     .

I am looking to fill in the gap between each status updates. The desired output table would look like the following:

ID Year Month Status
A  2014   1    Usub
A  2014   2    Usub
A  2014   3    Usub
A  2014   4    Sub
A  2014   5    Sub
A  2014   6    Sub
A  2014   7    Sub
A  2014   8    Usub
A  2014   9    Usub
A  2014  10    Usub
A  2014  11    Usub
A  2014  12    Usub
B  2014   1    Sub
B  2014   2    Sub
B  2014   3    Sub
B  2014   4    Sub
B  2014   5    Usub
B  2014   6    Usub
B  2014   7    Usub
B  2014   8    Usub
B  2014   9    Sub
B  2014  10    Sub
B  2014  11    Sub
B  2014  12    Sub
C  2014   1     .
.    .    .     .
.    .    .     .

Each ID has at least one status value. If an ID's first status record is "Usub", then all prior months' status is "Sub". (Like ID B on 2014/05) On the contrary, if the first status record start with "Sub", all prior months' status is "Usub"

Answer 1

You can generate an alternating sequence which is equivalent to your desired Status column by subtracting Status == "Usub" and Status = "Sub" in which way all positions that should be filled with Sub will have a lower value than those that should be filled with Usub and then use the fact that factor can be labelled in an ordered way to transform the integer sequence to a factor:

library(dplyr)
df %>% group_by(ID) %>% mutate(Status = factor(cumsum((Status == "Usub") - (Status == "Sub")), 
                                               labels = c("Sub", "Usub")))

#    ID Year Month Status
# 1   A 2014     1   Usub
# 2   A 2014     2   Usub
# 3   A 2014     3   Usub
# 4   A 2014     4    Sub
# 5   A 2014     5    Sub
# 6   A 2014     6    Sub
# 7   A 2014     7    Sub
# 8   A 2014     8   Usub
# 9   A 2014     9   Usub
# 10  A 2014    10   Usub
# 11  A 2014    11   Usub
# 12  A 2014    12   Usub
# 13  B 2014     1    Sub
# 14  B 2014     2    Sub
# 15  B 2014     3    Sub
# 16  B 2014     4    Sub
# 17  B 2014     5   Usub
# 18  B 2014     6   Usub
# 19  B 2014     7   Usub
# 20  B 2014     8   Usub
# 21  B 2014     9    Sub
# 22  B 2014    10    Sub
# 23  B 2014    11    Sub
# 24  B 2014    12    Sub

A corresponding data.table way would be:

library(data.table)
setDT(df)[, Status := as.character(factor(cumsum((Status == "Usub") - (Status == "Sub")), labels = c("Sub", "Usub"))), .(ID)]

where you have to convert the new factor back to character class since it doesn't allow the type change when creating a new column in place.

The data assumes you have empty string instead of NA :

structure(list(ID = c("A", "A", "A", "A", "A", "A", "A", "A", 
"A", "A", "A", "A", "B", "B", "B", "B", "B", "B", "B", "B", "B", 
"B", "B", "B"), Year = c("2014", "2014", "2014", "2014", "2014", 
"2014", "2014", "2014", "2014", "2014", "2014", "2014", "2014", 
"2014", "2014", "2014", "2014", "2014", "2014", "2014", "2014", 
"2014", "2014", "2014"), Month = c("1", "2", "3", "4", "5", "6", 
"7", "8", "9", "10", "11", "12", "1", "2", "3", "4", "5", "6", 
"7", "8", "9", "10", "11", "12"), Status = c("", "", "", "Sub", 
"", "", "", "Usub", "", "", "", "", "", "", "", "", "Usub", "", 
"", "", "Sub", "", "", "")), .Names = c("ID", "Year", "Month", 
"Status"), row.names = c(NA, 24L), class = "data.frame")

Answer 2

uniquevector<-unique(dat$ID)
for(i in uniquevector){
  zzz <- which(dat$ID==i & dat$Status == "Sub")
  zzz2 <- which(dat$ID==i & dat$Status == "Usub")
  zzz3 <- which(dat$ID==i & dat$Month == 12)
  zzz4 <- which(dat$ID==i & dat$Month == 1)
  if(zzz2 > zzz){  
  index<-zzz:(zzz2-1)
  dat$Status[index] <- "Sub"
}
  if(zzz2 < zzz){  
    index<-zzz2:(zzz-1)
    dat$Status[index] <- "Usub"
  }
  if(zzz3 > zzz2 & zzz < zzz2){  
    index<-zzz2:zzz3
    dat$Status[index] <- "Usub"
  }
if(zzz2 < zzz & zzz3 > zzz){  
  index<-zzz:zzz3
  dat$Status[index] <- "Sub"
if((zzz4 < zzz) & zzz < zzz2){  
    index<-zzz4:(zzz-1)
    dat$Status[index] <- "Usub"
}
  if((zzz4 < zzz2) & zzz2 < zzz){  
    index<-zzz4:(zzz2-1)
    dat$Status[index] <- "Sub"
  } 

  }}

Answer 3

Another option would be to convert the blank "" to NA and make use of na.locf from zoo package to replace the NA with the non-NA previous element. As this is a group by operation, we can also do this with ave from base R .

library(zoo)
df$Status <- with(df, ave(replace(Status, !nzchar(Status), NA), ID,
           FUN = function(x){ x1 <- na.locf(x, na.rm=FALSE)
         replace(x1, is.na(x1), setdiff(unique(na.omit(x1)), x1[!is.na(x1)][1]))}))
df$Status
#[1] "Usub" "Usub" "Usub" "Sub"  "Sub"  "Sub"  "Sub"  "Usub" "Usub" "Usub" "Usub" "Usub" "Sub"  "Sub"  "Sub"  "Sub"  "Usub" "Usub" "Usub"
#[20] "Usub" "Sub"  "Sub"  "Sub"  "Sub"