String replace regex, removing symbols from facebook content

Question

Hello I'm wondering how to filter out these custom Facebook symbols.

CULTURE CLUB PRESENTS JENNIFER CARDINI
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬

We proudly presents JENNIFER CARDINI
Support by Monsieur Moustache & Thang



▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
LINE-UP
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬

#TOWNHALL

▮ Monsieur Moustache

▮ Thang

▮ JENNIFER CARDINI


▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
JENNIFER CARDI ...

The expected result is the same text without the thick hyphens and music symbols.

I have very little knowledge about regex. I'm using this simple function but I think I need another replacement pattern.

 function zonderRareSymbolen(str) {
   new_string = str.replace(/^([ A-Za-z0-9_@#+-.'"]+(\r)?(\n)?)*$/g, ''); 
   return new_string.toLowerCase();
 }

Is it possible? I already tried litteraly putting pasting the symbols in the replace pattern. Can't I use a pattern that only allows letters and numbers?

UPDATE This function works if I literally put the text in:

zonderRareSymbolen('... Kisses ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ Come ...');
function zonderRareSymbolen(str) {
  new_string = str.replace(/▬/g, ''); 
  return new_string;
 }

But not on my detail.description value from an object:

zonderRareSymbolen(String(detail.desciption));
function zonderRareSymbolen(str) {
  new_string = str.replace(/▬/g, ''); 
  return new_string;
 }

Any ideas?

UPDATE The replace function doesn't seem to work with the custom symbol. All I can do now is work around it and only allow the desired characters like this:

function zonderRareSymbolen(str) { return (str).replace(/[^a-zA-Z0-9- ,=?&:.'"@/]/g, ' '); }

SOLVED By Mariano. After putting in <meta charset="UTF-8"> I am able to litterally paste the symbols into my replace function and replace them.

The subject string is the text in the image ... the expected result is the same text without the thick hyphens and music symbols — Auguste Van Nieuwenhuyzen, Nov 21 '15 at 02:18
Exactly, so if someone reading your post wants to copy that thick hyphen to include in a regex, he should just guess what's in the image? What if there are html tags you're not showing? What about non-breaking spaces? Where are the newlines? Think about it. You're asking people to help you parse text and you don't even give them the original text. — Mariano, Nov 21 '15 at 02:24
Thanks for adding the subject. Notice I **[edited](http://stackoverflow.com/posts/33789665/revisions)** the post to format the text as code... You mentioned you already tried using the char as literal in the pattern. Did you try `str.replace( /▬+\r?\n?/g, '').toLowerCase();` ? — Mariano, Nov 21 '15 at 02:51

score 1 · Accepted Answer · edited May 23 '17 at 10:27

1

You're not assigning the replaced value to detail.description. Use one of these options:

Assign the returned value to your object property:

    var detail = {
        id: 'some id',
        description: ".... Kisses\n▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬\nCome ...."
    };
    function zonderRareSymbolen(str) {
        return str.replace(/\u25AC+\r?\n?/g, '');
    }

    //asign the returned value
    detail.description = zonderRareSymbolen(detail.description);

    //print object
    result.innerText = JSON.stringify(detail);

    <body><pre id="result"></pre></body>

Or call by sharing, referencing your object property in the function context:

    var detail = {
        id: 'some id',
        description: ".... Kisses\n▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬\nCome ...."
    };
    function zonderRareSymbolen(obj) {
        //obj is a reference to the object passed by value
        obj.description = obj.description.replace(/\u25AC+\r?\n?/g, '');
    }

    //call by sharing
    zonderRareSymbolen(detail);

    //print object
    result.innerText = JSON.stringify(detail);

<body><pre id="result"></pre></body>

JavaScript passes variables by value. But an object's value is in fact a reference, thus, within the function context, a change in one of its members will persist. For detailed info, see Is JavaScript a pass-by-reference or pass-by-value language?.

edited May 23 '17 at 10:27

Community

1
1

answered Nov 23 '15 at 13:28

Mariano

6,423
4
31
47

I appreciate your effort a lot! Still I litterally tested every possibillity with your code but it does not affect my object. Strange that it doesn't work on my fetched object but it does on a static object. I figured out it the replace doesn't want to work with the unknown symbols ... so I'd rather use an approach like this `(detail.description).replace(/[^a-zA-Z-:. ]/g, ' ')` – Auguste Van Nieuwenhuyzen Nov 24 '15 at 07:24
@Auguste This is not related to `replace` or the regex. If it works for you here in the code snippet, it works in your script. Probably, `detail.description` has not been set yet when you call this function (is it being set with an `asynchronous` call?). I suggest checking for `if (detail.description === undefined)` before calling this function and debug it from there. But there isn't much I can do, since I don't have your code to understand what's going on... And if you can't debug it, I'd ask a new question, showing the relevant part of your code. Concentrate on `detail`, not the `replace` – Mariano Nov 24 '15 at 07:41
I checked it with a semaphore and I execute if the calls are completed, I know it's strange that it perfectly works here. If you really want I can send you the code, I tried everything. Whenever I change the replace pattern, it works, just not for the stripes. So I'm going to use the work around and only allow certain characters. – Auguste Van Nieuwenhuyzen Nov 24 '15 at 08:08
I also noticed that it only works for the detail object not for it's properties? I don't need the whole object to be stringified. – Auguste Van Nieuwenhuyzen Nov 24 '15 at 08:11
@Auguste I can imagine 2 possibilities: **1.** The *stripes* are a different char than what you pasted in the question (unlikely), or **2.** `detail.description` is undefined when you call it or `.description` is not a property, but a method... Since my solution is working for the particular case, I suggest asking a new question and let the community check your what's wrong, not just me. – Mariano Nov 24 '15 at 08:17
@Auguste `JSON.stringify` was an easy way to print the whole object here, just for testing purposes... Of course, in your script, you won't use that, you'd just refer to `detail.description` where needed – Mariano Nov 24 '15 at 08:19
I added a picture with the console output, the replace function doesn't seem to work with that symbol. If I put this `.replace(/ab/]/g, '**')`, it perfectly executes. – Auguste Van Nieuwenhuyzen Nov 24 '15 at 08:38
I don't know either but I have a work around so that's great for me, I appreciate all the time you put in it and gave me some valuable knowledge, thank you – Auguste Van Nieuwenhuyzen Nov 24 '15 at 09:11
@Auguste One last test: try `.replace(/\u25AC+\r?\n?/g, '')` – Mariano Nov 24 '15 at 09:22
That fixes it, thanks but there are so many characters I have to get out and it seems to take time to find the escaped values. At the moment I'm using this, `.replace(/[^a-zA-Z0-9- ,=?&:.'"@\u00C1\u00C0\u00C2\u00C4\u00C7\u00C8\u00C9\u00CA\u00CB\u00CC\u00CD\u00CE\u00CF\u00D1\u00D2\u00D3\u00D4\u00D5\u00D6\u00D9\u00DA\u00DB\u00DC\u00E1\u00E0\u00E2\u00E4\u00E7\u00E8\u00E9\u00EA\u00EB\u00EC\u00ED\u00EE\u00EF\u00F1\u00F2\u00F3\u00F4\u00F5\u00F6\u00F9\u00FA\u00FB\u00FC\/+\r?\n?]/g, '');` It works fine except for the `\r\n` chars. What you posted, filters that specific character. Thank you for that! – Auguste Van Nieuwenhuyzen Nov 30 '15 at 14:29
@AugusteVanNieuwenhuyzen aha! you couldn't paste the character because of your source encoding. Make sure you save your JS file as UTF-8 and define the same charset in HTTP headers (or wherever you're running this). You could simplify it using ranges in regex, e.g. `[\u0080-\u00FF]`... Or another approach: define the chars you want to keep, e.g. `[^\u0020-\u007E]` – Mariano Dec 01 '15 at 11:55
Yes my bad so I just put in ``?? :p Thank you for everything, helped me a lot for my end terms! – Auguste Van Nieuwenhuyzen Dec 01 '15 at 12:33
@Auguste Yes, and set UTF-8 in your text editor too... And test it... If it doesn't work, you can always go back to the `\uNNNN` notation :) – Mariano Dec 01 '15 at 12:42
It works, I put it in my html pages. So now I can do stuff like this `.replace(/─/g, '').replace(/▔/g, '').replace(/◙/g, '');` however it doesn't seem to work if I do this `.replace(/▬─▔◙/g, '');`. How do I put that in my text editor? Using Sublime Text – Auguste Van Nieuwenhuyzen Dec 01 '15 at 12:49
`/▬─▔◙/g` matches that specific sequence. Use a **[character class](http://www.regular-expressions.info/charclass.html)** `/[▬─▔◙]+/g` – Mariano Dec 01 '15 at 12:52
Thank you very much! There's one last problem, this character is for new lines: `↵` I can't seem to replace it for `\r\n` Maybe because my editor doesn't recognize it or because it is invisible in html? Would you know a solution to that? You really saved me already, so no problem if you don't feel like it. – Auguste Van Nieuwenhuyzen Dec 01 '15 at 14:08

String replace regex, removing symbols from facebook content

1 Answers1