How to select all characters except some specific words using regex?

Question

Good morning to all,

I would like to run a string search for some specific strings, but in the project we are working on we have no access to any function find or similar, only to a regular expression-based substitution function (something like replace(string, regex, replacement)).

The idea then would be to select all the characters EXCEPT the sequences I want to find. Thus, I would remove these unwanted characters and compare with what I want to find.

Example (not a specific language):

string expReg = ??????;
string texto = "xxxxxxxxboloxxxxxxxfarinhaxxxxxxacucarxxxx";
string busca = replace(texto, expReg, "");
if(busca == "bolofarinhaacucar"){
    return("Sucesso");
}

Luckily the words we need to find need to be in the defined order, so it would not be necessary to include all permutations.

We try to find some solution using regular expressions, but we always bump into the problem that the positive lookbehind (?<=ABC) not supported in Javascript.

Any idea?

Answer 1

To remove a certain character from a sequence just make the following substitution (example in Javascript):

var texto = 'xxxxxxxxboloxxxxxxxfarinhaxxxxxxacucarxxxx';

var expReg = /([x]+)/g; // Vai procurar por uma ou mais ocorrências de "x"
var busca = texto.replace(expReg, '');

if(busca == "bolofarinhaacucar"){
    console.log("Sucesso");
}

DEMO

To deny the correspondence(match) of certain characters just use the character class denied using ^ at the beginning of class.

var texto = 'xxxxxxxxboloxxxxxxxfarinhaxxxxxxacucarxxxx';

var expReg = /([^x]+)/g;
var buscaArray = texto.match(expReg, '').toString(); // match retorna um array com os valores encontrados
var busca = buscaArray.split(",").join(""); // Transformamos para string para poder comparar

if(busca == "bolofarinhaacucar"){
    console.log("Sucesso");
}

DEMO

Maybe the syntax changes depending on the engine used. Here(in English) shows a comparison between the Engines of regular expressions.

Answer 2

it would be good to know the language you are working to know a more efficient way to help you since denying word in regex is not easy

http://aurelio.net/regex/guia/negar-palavra.html#5_3

As you said it will always be those specific words can do something like this

expReg = "\w+(bolo)\w+(farinha)\w+(acucar)\w+"

string texto = "xxxxxxxxboloxxxxxxxfarinhaxxxxxxacucarxxxx";
string busca = replace(texto, expReg, "$1$2$3"); //substitui pelo grupo 1, 2 e 3

-- Can test here: https://regex101.com/r/oP3kU3/1