script.r

#!/usr/bin/env Rscript

# Given:
#
# V | X1 | X2 | X3 | Y1 | Y2 | Y3
# -------------------------------
# A | A  | B  | C  | 2  | 3  | 5   <-- A matches X1,    return 2
# B | A  | B  | B  | 6  | 4  | 5   <-- B matches X2,X3, return 4,5
#
# We want:
#
# 2
# 4,5

df <- data.frame(
  "V" = c("A", "B"),
  "X1" = c("A", "A"),
  "X2" = c("B","B"),
  "X3" = c("C", "B"),
  "Y1" = c(2, 6),
  "Y2" = c(3, 4),
  "Y3" = c(5, 5)
)

print(df, row.names = FALSE)

fun <- function (row) {
  # This vector will collect all values to be returned.
  ret = c()
  # Extract value in V column.
  V <- row[1]
  # Get ID of the last X column assuming that:
  # - the first column is used by V
  # - there are as many X columns as Y columns
  # In our example lastXcol is equal to 4.
  lastXcol = (length(row)-1)/2+1
  # Loop over all X columns.
  for (colid in 2:lastXcol) {
    # Extract value of the current X column.
    X <- row[colid]
    # If value in X column == V
    if (X == V) {
      # Then extract value of the corresponding Y column.
      Y <- row[colid+lastXcol-1]
      # And put this value in our vector.
      ret <- append(ret, Y)
    }
  }
  # Return collected values, glued with a comma.
  return(paste(ret, collapse = ","))
}

df2 <- apply(df, 1, fun)

print(df2, row.names = FALSE)