具有多个列和NA的R嵌套Ifelse

Question

I have a dataframe for races:

df <- data.frame(ID=c(1,2,3,4,5,6), Condition=c(1,2,2,1,1,2), White=c(1,1,NA,NA,NA,NA), Black=c(2,NA,NA,NA,2,NA), Asian=c(NA, NA, NA, 3, 3, 3), AmerIndian=c(NA,NA,4,NA,NA,NA), NatHawaiian=c(NA, NA, NA, 5, NA, NA))

and I'd like to develop a new field for race where whatever Condition 2 is indicated a new field will be filled in.

this is what I tried:

df$var <-ifelse(as.numeric(df$White)==1&!is.na(df$White),"White", 
+     ifelse(as.numeric(df$Black)==2&!is.na(df$Black),"Black",
+        ifelse(as.numeric(df$Asian)==3&!is.na(df$Asian),"Asian",
+           ifelse(as.numeric(df$AmerIndian)==4&!is.na(df$AmerIndian),"AmerIndian",
+               ifelse(as.numeric(df$NatHawaiian)==5&!is.na(df$NatHawaiian),"NatHawaiian",NA)))))

and I received this error:

Error in +ifelse(as.numeric(df$NatHawaiian) == 5 & !is.na(df$NatHawaiian),  : 
  invalid argument to unary operator

Answer 1

Based on the code you input and the error generated, I suggest removing the plus symbols at the begining of each ifelse statement. When that is done, the code you shared:

df$var <- ifelse(
  as.numeric(df$White) == 1 & !is.na(df$White),
  "White", ifelse(
    as.numeric(df$Black) == 2 & !is.na(df$Black),
    "Black", ifelse(
      as.numeric(df$Asian) == 3 & !is.na(df$Asian),
      "Asian", ifelse(
        as.numeric(df$AmerIndian) == 4 & !is.na(df$AmerIndian),
        "AmerIndian", ifelse(
          as.numeric(df$NatHawaiian) == 5 &
            !is.na(df$NatHawaiian),
          "NatHawaiian",
          NA
        )
      )
    )
  )
)

Will generate this output:

  ID Condition White Black Asian AmerIndian NatHawaiian        var
1  1         1     1     2    NA         NA          NA      White
2  2         2     1    NA    NA         NA          NA      White
3  3         2    NA    NA    NA          4          NA AmerIndian
4  4         1    NA    NA     3         NA           5      Asian
5  5         1    NA     2     3         NA          NA      Black
6  6         2    NA    NA     3         NA          NA      Asian

Answer 2

you can try with tidyverse like:

df %>%
  gather(key = "race", value = "val", 3:7) %>%
  mutate( rc = if_else((Condition == 2 & !is.na(val)), race, NULL)) %>%
  spread(race, val)