CODEWITHSUNDEEP
rdataframe
rr_silva
rr_silva

Reputation: 54

Automatically fill a column in a dataframe in R with "coordinates"

Starting with the Dataframe y:

x <- c(2,NA,6,8,9,10)
y <- data.frame(letters[1:6], 1:6, NA, 3:8, NA, x, NA)
colnames(y) <- c("Patient", "C1", "First_C1", "C2", "First_C2", "C3", "First_C3")

I want R to look at each element of C1, find out the first patient (first row) with that element and the identify in which column it is, and add the "coordinates" "Patient_Column" to First_element_C1... Then, do the same with C2 and C3.

So, the result should be this:

y$First_C1 <- c("a_C1", "a_C3", "a_C2", "b_C2", "c_C2", "c_C3")
y$First_C2 <- c("a_C2", "b_C2", "c_C2", "c_C3", "e_C2", "d_C3")
y$First_C3 <- c("a_C3", NA, "c_C3", "d_C3", "e_C3", "f_C3")

I dont know how to write the code, not even how to search for it... Could someone help me here?

Upvotes: 0

Views: 367

Answers (1)

nicola
nicola

Reputation: 24480

We start from the y without the output columns:

y<-structure(list(Patient = structure(1:6, .Label = c("a", "b", 
"c", "d", "e", "f"), class = "factor"), C1 = 1:6, C2 = 3:8, C3 = c(2, 
NA, 6, 8, 9, 10)), .Names = c("Patient", "C1", "C2", "C3"), row.names = c(NA, 
-6L), class = "data.frame")

Then, we can try:

y[paste0("First_C",1:3)]<-lapply(y[,2:4],
      function(x) {
         d<-arrayInd(match(x,t(y[,2:4])),dim(t(y[,2:4])))[,2:1]
         paste(y$Patient[d[,1]],colnames(y[,2:4])[d[,2]],sep="_")
})
y[,5:7][is.na(y[,2:4])]<-NA
#  Patient C1 C2 C3 First_C1 First_C2 First_C3
#1       a  1  3  2     a_C1     a_C2     a_C3
#2       b  2  4 NA     a_C3     b_C2     <NA>
#3       c  3  5  6     a_C2     c_C2     c_C3
#4       d  4  6  8     b_C2     c_C3     d_C3
#5       e  5  7  9     c_C2     e_C2     e_C3
#6       f  6  8 10     c_C3     d_C3     f_C3

Upvotes: 2

Related Questions