Finding ALL indices of positions in vector matching columns of DF in R

Question

I have a dataframe A whose columns I want to match with the row.names of another dataframe B.

# A
v1    v2    
X1    X3
X1    X5
X1    X15
X2    X3
X2    X4
...


# row.names of B (some values are duplicated)
row_names_B=c('X17', 'X1', 'X2', 'X15', 'X3', 'X3', 'X1', 'X5', 'X4', ...)

I want to match the columns of A with the positions of row_names_B, such that I can return a list of ALL positions in B for each row in A.

#my results:
v1_index    v2_index
2           5        #matches X1 in pos 2, X3 in pos 5
2           6        #matches X1 in pos 2, X3 in pos 6
7           5        #matches X1 in pos 7, X3 in pos 5
7           6        #matches X1 in pos 7, X3 in pos 6
2           5        #matches X1 in pos 2, X3 in pos 8
7           5        #matches X1 in pos 7, X3 in pos 8
...

Note that I want to find all possible solutions.

I understand that this should be with some variant of match or which as given in this example, but I'm not sure how to do the explosion for each of the matches. The way I see it is by running it through for loops, row by row, but perhaps there is a better way to do this?

Answer 1

You could create a list of position based on their name and randomly assign one value in the dataframe A from the list of positions.

C <- A
ref <- split(seq_along(row_names_B), row_names_B)
C[] <- lapply(A, function(y) sapply(ref[y], 
                    function(x) if(length(x) == 1) x else sample(x, 1)))
C

#  v1 v2
#1  2  5
#2  2  8
#3  7  4
#4  3  5
#5  3  9

data

A <- structure(list(v1 = c("X1", "X1", "X1", "X2", "X2"), v2 = c("X3", 
"X5", "X15", "X3", "X4")), class = "data.frame", row.names = c(NA, -5L))
row_names_B <- c("X17", "X1", "X2", "X15", "X3", "X3", "X1", "X5", "X4")