# [R] Removal of columns from matrix where all values of the column are identical.

Benjamin Ward (ENV) B.Ward at uea.ac.uk
Sat Jan 26 14:37:40 CET 2013

```Hi,

I've been trying to work this out how it works, I'm still not totally sure but seems to me it subsets the matrix according to whether each column returns all TRUE, to not being the same values when you compare the column[-1] and the column[-length(column)], essentially siding the column against itself and making comparison? It seems that it also removed columns with any repeated values, rather than columns in which all values are the same:

testm<-matrix(nrow=5, ncol=5)
testm[,1] <- c(1,2,3,4,5)
testm[,2] <- c(3,3,3,3,3)
testm[,3] <- c(3,3,3,4,3)
testm[,4] <- c(5,4,3,2,1)
testm[,5] <- c(1,2,3,4,4)
testm

test3 <- testm[,apply(testm,2,function(x) all(c(TRUE,x[-length(x)]!=x[-1])))]
test3

test3
[,1] [,2]
[1,]    1    5
[2,]    2    4
[3,]    3    3
[4,]    4    2
[5,]    5    1

Thanks,
Ben W.

________________________________________
From: arun [smartpink111 at yahoo.com]
Sent: 26 January 2013 02:34
To: Benjamin Ward (ENV)
Cc: R help
Subject: Re: [R] Removal of columns from matrix where all values of the column are identical.

Hi,

I guess this should also work:

Matrix[,apply(Matrix,2,function(x) all(c(TRUE,x[-length(x)]!=x[-1])))]
#     [,1] [,2] [,3]
#[1,]    1    5    5
#[2,]    2    4    1
#[3,]    3    3    4
#[4,]    4    2    3
#[5,]    5    1    2
A.K.

----- Original Message -----
From: Benjamin Ward (ENV) <B.Ward at uea.ac.uk>
To: "r-help at r-project.org" <r-help at r-project.org>
Cc:
Sent: Friday, January 25, 2013 6:17 PM
Subject: [R] Removal of columns from matrix where all values of the column are identical.

Hi all,

I'd like to write a piece of code which will remove columns from a matrix, if the column contains only one value, say every value in the column is a "3":

Matrix <- matrix(NA, nrow=5, ncol=4)
Matrix[,1] <- c(1,2,3,4,5)
Matrix[,2] <- c(3,3,3,3,3)
Matrix[,3] <- c(5,4,3,2,1)
Matrix[,4] <- c(5,1,4,3,2)

[,1] [,2] [,3] [,4]
[1,]    1    3    5    5
[2,]    2    3    4    1
[3,]    3    3    3    4
[4,]    4    3    2    3
[5,]    5    3    1    2

What I have written so far is a loop which will see if all values are the same, a bit of a hack since it just checks all values are equal to the first value of the column, if not, by definition the column cannot contain only one value/variable/character:

removals<-c()
for(i in 1:ncol(Matrix)){
if(all(Matrix[,i] == Matrix[[1,i]])){
removals<-append(removals, i)
}
}
new.Matrix <- Matrix[,-removals]

This works for matrices with numbers or characters.
My question is - is there a better or more efficient way of doing this, maybe with apply or something. My first thought was apply set to operate over all columns, but was unsure of the indexing and selecting columns to be deleted.

Thanks,

Ben W.

University of East Anglia (ENV): b.ward at uea.ac.uk
The Sainsbury Laboratory: ben.ward at sainsbury-laboratory.ac.uk

[[alternative HTML version deleted]]

______________________________________________
R-help at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help