Speed up a function that checks for a sequence

Question

I created a function yes.seq that takes two arguments, a pattern pat and data dat, the function looks for the presence of a pattern in the data and in the same sequence.

For example:

dat <- letters[1:10]
dat
[1] "a" "b" "c" "d" "e" "f" "g" "h" "i" "j"
pat <- c('a',"c","g")
 
 yes.seq(pat = pat,dat = dat)
[1] TRUE

because this sequence is in the pattern and in the same order.

"a" "b" "c" "d" "e" "f" "g" "h" "i" "j"

If, for example, we reverse the pattern, then we get FALSE:

yes.seq(pat = pat,dat = **rev(dat)** )
[1] FALSE

Here is my function:

yes.seq <- function(pat , dat){ 
 lv <- rep(F,length(pat))
 k <- 1 
 for(i in 1:length(dat)){ 
 if(dat[i] == pat[k]) 
 {
 lv[k] <- TRUE
 k <- k+1 
 } 
 if(k==length(pat)+1) break
 }
 return( all(lv) )
}

I am not satisfied with the speed of this function. Can you help me with that?

• minem: can you add example data on which the function doesn't perform well?
• Henrik: Cross-posted on SO: Match all elements of a pattern with a vector and in the same order. Please read Is it OK to cross post here and in SO? "No, it's not ok. Yes, it would be bad manners."

Answer 1

Here is a vectorized version:

yes.seq <- function(dat, pat) {
 paste(dat[dat %in% pat], collapse = "") == paste(pat, collapse = "")
}
yes.seq(dat, pat)
# [1] TRUE
yes.seq(dat, rev(pat))
# [1] FALSE