A smarter solution to matching many to many

better.R

# For each element, match against the data frame of longer vectors
compSlightlySmarter <- function(element, table) {
  matches <- sapply(table, function(x) {
    # finds exact matches of element to the 'table' (one column in the longer df)
    locations <- match(element, as.character(x))
    # if there are no unbroken or unordered sequences, it's a match!
    return(!any(is.na(locations)) & !is.unsorted(locations))
  })
  if(any(matches)) {
    # unlike the dumb version, we return matches
    return(which(matches))
  }
}

# a list of either NULL values for no matches or the column index of the table dataframe
lapply(shrt, compSlightlySmarter, table = long)

Attempting to run this over a df produces (unsurprisingly) an error. This'll be interesting ;p.