将两列合并为R中的一列

Question

V1<-c("Apple", "Orange", "NA", "NA")
V2<-c("NA", NA", "Strawberries", "NA")
V3<-data.frame(V1, V2)

Want to get the following:

V4<-c("Apple", "Orange", "Strawberries", "NA")
V4<-data.frame(V4)

Any suggestions? Thanks!!

Answer 1

V1<-c("Apple", "Orange", "NA", "NA")
V2<-c("NA", "NA", "Strawberries", "NA")
V3<-data.frame(V1, V2)


V3$V4 <- with(V3, ifelse(V1 == "NA" & V2 != "NA", as.character(V2), 
                         ifelse(V1 != "NA" & V2 == "NA", as.character(V1), "NA")))

This gives:

      V1           V2           V4
1  Apple           NA        Apple
2 Orange           NA       Orange
3     NA Strawberries Strawberries
4     NA           NA           NA

Answer 2

Using dplyr 's coalesce , also reading in the strings not as factors:

V1<-c("Apple", "Orange", NA, NA)
V2<-c(NA, NA, "Strawberries", NA)
V3 <-data.frame(V1, V2, stringsAsFactors = FALSE)

library(dplyr)
V3 %>% mutate(V4 = coalesce(V1, V2))

      V1           V2           V4
1  Apple           NA        Apple
2 Orange         <NA>       Orange
3   <NA> Strawberries Strawberries
4   <NA>         <NA>         <NA>

Answer 3

There is also the possibility to use case_when from dplyr. I like it because it is much easier to keep track of what is happening compared to nested ifelse clauses:

library(dplyr)

V1<-c("Apple", "Orange", "NA", "NA")
V2<-c("NA", "NA", "Strawberries", "NA")
V3<-data.frame(V1, V2)


V3 <- V3 %>%
mutate(V4 = case_when(V1 == "NA" & V2 != "NA" ~ as.character(V2),
                    V1 != "NA" & V2 == "NA" ~ as.character(V1),
                    TRUE ~ as.character("NA") ) )

In this case the difference is probably not so big, but the more conditions are tested, the easier it gets. If for instance you also need to treat the case where there is a non-"NA" value in both columns, this would simply extend to something like:

V1<-c("Apple", "Orange", "NA", "NA", "Banana")
V2<-c("NA", "NA", "Strawberries", "NA", "Pear")
V3<-data.frame(V1, V2)


V3 <- V3 %>%
mutate(V4 = case_when(V1 == "NA" & V2 != "NA" ~ as.character(V2),
                    V1 != "NA" & V2 == "NA" ~ as.character(V1),
                    V1 != "NA" & V2 != "NA" ~ paste(V1,V2),
                    TRUE ~ as.character("NA")  ) )