Apply a function to elements of matrix on condition - r

Question

I'm looking to change the value of a certain entry in a matrix based on the value of another entry. Its easiest to explain with an example:
Matrix
ABC-DEF 1 0 0 0
HIJ-KLM 0 0 0 0
NOP-QRS 1 0 0 0
KLM-HIJ 0 0 0 0
DEF-ABC 0 0 0 0
QRS-NOP 0 0 0 0
As you can see, each of the rows in the matrix above has a counterpart (e.g. ABC-DEF's counterpart is DEF-ABC).
Is there some way in which I can look to see which rows have a one in the first column and then place a 2 in the fourth column of its counterpart? In the above example then:
ABC-DEF 1 0 0 0
HIJ-KLM 0 0 0 0
NOP-QRS 1 0 0 0
KLM-HIJ 0 0 0 0
DEF-ABC 0 0 0 2
QRS-NOP 0 0 0 2
I'm quite stuck and would really appreciate any help!
Thanks!

Answer 1

Assuming your column names are V1,...,V5, you can do something like this :
values <- d$V1[d$V2==1]
d$V5[d$V1 %in% gsub("(...)-(...)","\\2-\\1", values)] <- 2
Which will give :
V1 V2 V3 V4 V5
1 ABC-DEF 1 0 0 0
2 HIJ-KLM 0 0 0 0
3 NOP-QRS 1 0 0 0
4 KLM-HIJ 0 0 0 0
5 DEF-ABC 0 0 0 2
6 QRS-NOP 0 0 0 2
If, instead of a data frame, your data is a numeric matrix m with row names, you can do :
values <- rownames(m)[m[,1]==1]
m[rownames(m) %in% gsub("(...)-(...)","\\2-\\1", values),4] <- 2
EDIT : To understand what the code is doing, you must see that :
gsub("(...)-(...)","\\2-\\1", values)
will replace any character string in the values vector of the form XXX-YYY by YYY-XXX via regexp matching. The result is a character vector of the "counterparts" of values. Then we use %in% to select every rows whose rownames appear in these counterpart values, and assign 2 in the fourth column for these rows.