Efficiency considerations.

• Try not to work with too many regular expressions at once. Before a regular expression can be used, it must be compiled (which re_match/5 does automatically). re_match/5 maintains a cache of compiled regular expressions, so they do not need to be compiled each time they are used. However, if more than 8-10 expressions are used simultaneously, repeated recompilation might result.
• When a list of characters is passed to any one of the above predicates, it is converted into a C string. This can be expensive, if done too often for the same string.

  One way to circumvent the problem is to use atom_codes/2 to first convert the list into an atom and then use that atom repeatedly in the match operations. One problem here might be the aforementioned overflow of the atom table. So, if this is a concern, the following predicate (which always succeeds) can help:

     re_charlist_to_string(+ListOfCharacters, -String)
  

  This predicate converts lists of characters into uninterned strings, which can be used without the fear of atom table overflow:

  | ?- re_charlist_to_string("abcdefg",L).
  
  L = abcdefg
  

  The resulting string can be passed to re_match/5, re_substitute/5, and re_substring/4 for further processing.

  Note, however: you cannot call re_charlist_to_string before you finished working with the string generated by the previous call: all calls to this function use the same static buffer to hold the output string, so each subsequent call to re_charlist_to_string will override the previously generated strings.