CODEWITHSUNDEEP
r
123hwc
123hwc

Reputation: 31

subset rows used in which.max in R

I am trying to get the max within a row, for a particular subset of rows.

a <- data.frame(a = rnorm(5), b = rnorm(5), c = rnorm(5), x = rnorm(5), y = rnorm(5), z = rnorm(5))

a
        a           b          c           x           y           z
1 -0.99415010  0.06177776  1.1224987 -0.60239594 -0.53526479  1.26826763
2 -1.17727409 -0.60488488  1.6509299  0.65014481 -0.91940395  0.47520214
3 -0.04944379  1.47238123 -1.1691980  0.07603221 -0.05844166 -0.24809870
4 -0.05230395 -2.52972650 -0.1170329  0.57833492  0.61692175 -0.16939056
5 -0.61666540  0.02606907  0.2438158 -1.01048883 -1.23118019  0.09784697

> names(a)[apply(a, 1, which.max)]
[1] "z" "c" "b" "y" "c"

Is there a way to subset the data so that I only get which.max results for columns x, y, and z? When I try to subset the columns I need, I get this error:

Error in names(---)[apply(----[, 13:68], 1, which.max)] : 
  invalid subscript type 'list'

Thank you!

Upvotes: 0

Views: 573

Answers (3)

akrun
akrun

Reputation: 886948

You could also try:

   b1 <- a[,letters[24:26]]
    names(b1)[max.col(b1)]
   #[1] "x" "z" "y" "y" "y"

data

   set.seed(42)
    a <- data.frame(a = rnorm(5), b = rnorm(5), c = rnorm(5), x = rnorm(5), y = rnorm(5), z = rnorm(5))

Upvotes: 0

ndr
ndr

Reputation: 1437

How about

 b <- a[,c("x","y","z")]
 names(b)[apply(b, 1, which.max)]

or

names(a[,4:6])[apply(a[,4:6], 1, which.max)]

Upvotes: 1

rnso
rnso

Reputation: 24535

Following may help (my 'a' values are different from yours since you did not use set.seed):

xx = apply(a, 1, which.max)
names(a)[xx[xx>3]]
[1] "y"

If you want to analyze only x,y,z columns:

> a = a[,4:6]
> names(a)[apply(a, 1, which.max)]
[1] "x" "y" "x" "z" "y"

Upvotes: 1

Related Questions