Conditional statement in R dataframe - r

Question

I have dataframe df as below.
dput(df)
structure(list(X = c(1, 2, 5, 7, 8), Y = c(3, 5, 8, 7, 2), Z = c(2,
8, 7, 4, 3), R = c(6, 6, 6, 6, 66)), .Names = c("X", "Y", "Z",
"R"), row.names = c(NA, -5L), class = "data.frame")
df
class(df)
I have to modify df under two conditions.
First:
modify df so that it check minimum between X,Y,Z for each row and whichever is minimum get replaced with corresponding value of R.
Second case:
which is minimum between X,Y,Z,R in each row, it get replaced with maximum between X,Y,Z,and R and create a new df.
How should i get that?
I tried ifelse and if and else but could not get what i want..
Any help would be appreciated.

Answer 1

You can create a new dataset "df1" with first three coumns of "df". Multiply "df1" with "-1" so that maximum values become "min" (assuming that there are no negative values). Here, in the example, the values were all unique per row. So, you can use the function max.col and specify the ties.method='first'. It will get you the index of maximum value (here it will be minimum) per row, cbind it will the 1:nrow(df) to create the "row/column" index and extract the elements of "df1" based on that index (df1[cbind..]) and change those values to "R" column values (<- df$R). You could then change the original "df" columns ("df[1:3]") to new values. If there are more than one "minimum" value per row, you could use the "loop" method described for the second case.
df1 <- df[1:3]
df1[cbind(1:nrow(df),max.col(-1*df1, 'first'))] <- df$R
df[1:3] <- df1
df
# X Y Z R
#1 6 3 2 6
#2 6 5 8 6
#3 6 8 7 6
#4 7 7 6 6
#5 8 66 3 66
Create a copy of "df" (df2), get the max values per row using pmax, loop over the rows of "df2" (sapply(seq_len...)) and change the "minimum" values in each row to corresponding "max" values ("MaxV"), transpose (t) and assign it back to the "df2" (df2[])
df2 <- df
#only use this if there is only a single "minimum" value per row
# and no negative values in the data
#df2[cbind(1:nrow(df), max.col(-1*df2, 'first'))] <-
# do.call(pmax, df2)
MaxV <- do.call(pmax, df2)
df2 [] <- t(sapply(seq_len(nrow(df2)), function(i) {
x <- unlist(df2[i,])
ifelse(x==min(x), MaxV[i], x)}))
df2
# X Y Z R
#1 6 3 6 6
#2 6 8 8 6
#3 8 8 7 8
#4 7 7 7 7
#5 8 66 66 66