1

I have this chunk of code in an R script, and what it does is that from a data frame (that has say 4 columns) finds and counts for each row the column with the maximal value. But when in a row there are ties (say two or maybe three max values), the code below just considers the first one (does not count the other ones). Is there in R any way to count the columns with max value even when there are multiple max values (ties?) in rows.

m_0 <- read.csv(file="Tests/myResults", head=FALSE, sep=",")

varb_m0 <- c(m0$ V4)

#create dataframe

myDataFrame <- data.frame(mode_0=varb_m0)

#find max
result <- apply(myDataFrame,1,which.max)

#factor it
result <- factor(result)

#print result
names(myDataFrame)[result]

summary(result)
4

1 に答える 1

2

A simple way to do it would be to make a function first: 

all_max <- function(x) {which(x == max(x))}  #Index of each value that is maximum.

And then just apply that to your data frame: 

result <- apply(myDataFrame,1,all_max)

I think the rest should be the same, though you didn't provide any data to test it on.

于 2012-08-03T15:48:25.387 に答える