Remove the rows in matrix A, which appear in the matrix B(B is a subset of A)

Question

I have a matrix A:

> A
      [,1] [,2] [,3]
[1,]    1    2    3
[2,]    4    5    6
[3,]    7    8    9
[4,]    1    2    3
[5,]    4    5    6
[6,]    7    8    9

and a matrix B:

> B
      [,1] [,2] [,3]
[1,]    1    2    3
[2,]    4    5    6

I want to delete the rows in matrix A, which appear in B. And the result should be:

> C
      [,1] [,2] [,3]    
[1,]    7    8    9
[2,]    7    8    9

Thanks a lot!

This should help: http://stackoverflow.com/questions/10907359/r-remove-rows-from-one-data-frame-that-are-in-another — Gaurav Taneja, Apr 13 '16 at 11:17
You could do sometype of an anti-join- see [here](http://stackoverflow.com/questions/28702960/find-complement-of-a-data-frame-anti-join). Using `data.table`, this could be `library(data.table);DTA <- data.table(A);DTB <- data.table(B);DTA[!DTB, on = names(DTA)]` — David Arenburg, Apr 13 '16 at 11:31
@GauravTaneja it works and it's a little tricky, thanks a lot — Ghrua, Apr 13 '16 at 11:39
@HuaYoung it actually doesn't work for your example. Simply because you have two duplicated rows in your first matrix which doesn't appear in your second matrix and hence also got removed. — David Arenburg, Apr 13 '16 at 11:46
@DavidArenburg you are right, I think I should update my problem. Actually I don't care the duplicated rows. Thank you! — Ghrua, Apr 13 '16 at 12:13
@DavidArenburg I am sorry that I cannot update the example, because some people have answered me. The rows of the real data that I am using are unique. — Ghrua, Apr 13 '16 at 12:36

score 2 · Accepted Answer · answered Apr 13 '16 at 12:31

2

We can use anti_join

library(dplyr)
as.matrix(anti_join(as.data.frame(A), as.data.frame(B)))
#     V1 V2 V3
#[1,]  7  8  9
#[2,]  7  8  9

answered Apr 13 '16 at 12:31

akrun

874,273
37
540
662

score 1 · Answer 2 · answered Apr 13 '16 at 11:20

1

First create the above two matricies

A <- matrix(rep(c(1:9),2),ncol=3,byrow = T)
B <- matrix(1:6,ncol=3,byrow = T)

Get the index of rows to be removed

Indx <- apply(B,1,function(x) apply(A,1,function(y) all.equal(x,y)))
Indx <- apply(Indx,1,any)

Remove the rows from first matrix

A[-which(Indx),]

answered Apr 13 '16 at 11:20

Koundy

5,265
3
24
37

It works for me. Although there are some warning messages, it doesn't matter. – Ghrua Apr 13 '16 at 12:26
See @akrun answer. Its perfect. You can also explore other join operations in dplyr package. – Koundy Apr 13 '16 at 12:51

score 0 · Answer 3 · answered Apr 13 '16 at 11:31

Make use of match_df function inside plyr package.

    library(plyr)
    A = matrix(rep(1:9,2),nrow=6,byrow=T)
    B = matrix(1:6,nrow=2,byrow=T)

    M_ROWS = as.numeric(rownames(match_df(data.frame(A),data.frame(B))))
    # Matching on: X1, X2, X3

    Result <- A[-M_ROWS,]
    Result
    #      [,1] [,2] [,3]
    # [1,]    7    8    9
    # [2,]    7    8    9

Remove the rows in matrix A, which appear in the matrix B(B is a subset of A)

3 Answers3