0
\$\begingroup\$

This question is kinda similar to mine. However, I am using C++ with Qt instead of C#.

How would I efficiently and easily remove all accents and special characters like !"§$%&/()=? etc. from a QString?

So "áche" should turn into "ache" or "über dir" to "ueber dir" (in german ü,ä,ö can be changed into the normalized character with an e appended) or at least "uber dir".

Note: Some people use a $ instead of s in some words so I want to make sure if a file is called "Ke$ha" that it will come out as "Kesha" or at least "KeSha".

The way I do it so far, incomplete, is like this:

void Utils::replaceInvalidChars(QString &str)
{
    if( str.size() == 0 )
        return;

    while( str.at(0) == '.' ) {
        str.remove(0,1);
    }

    str.replace( "/", "-" );
    str.replace( "|", "" );
    str.replace( ":", "-" );
    str.replace("\"", "" );
    str.replace( "?", "" );
    str.replace( "$", "s" );
    str.replace( "*", "" );
    str.replace( ",", "" );
    str.replace( "¿", "" );
    str.replace( "¡", "" );
    str.replace( "!", "" );
    str.replace( "'", "" );

    str.replace( "ë", "e" );
    str.replace( "ê", "e" );
    str.replace( "é", "e" );
    str.replace( "è", "e" );

    str.replace( "ç", "c" );

    str.replace( "ó", "o" );
    str.replace( "ö", "oe" );

    //U's...
    str.replace( "ü", "ue" );
    str.replace( "Ü", "U" );
    str.replace( "ù", "u" );
    str.replace( "Ù", "U" );
    str.replace( "û", "u" );
    str.replace( "Û", "u" );


    //ns
    str.replace( "ñ", "n" );

    //as
    str.replace( "ä", "ae" );
    str.replace( "Ä", "ae" );
    str.replace( "á", "a" );
    str.replace( "Á", "A" );
    str.replace( "à", "a" );
    str.replace( "À", "A" );

    str.replace( "ï", "i" );
}

So at first I remove all dots from the beginning. No matter how many there are. Then I replace certain characters with no character at all and some with a character like 's' or a depending on what it is.

My way is very long, tedious and chaotic. I am about to organize it a little with comments like "N's", "U's" etc. but still, if I make a mistake somewhere it will take way too long until I (eventually) find it.

\$\endgroup\$
3
\$\begingroup\$

I would start by separating the data from the logic:

std::vector<std::pair<QString, QString>> replacements { 
    { "/", "-" },
    { "|", ""  },
    // ...
    { "ï", "i" }
};

for ( auto const &r : replacements) { 
    str.replace(r.first, r.second);
}

I'm not sure the comments about the groups of letters being replaced really add a lot though.

Then I'd at least consider moving the data out of the program itself, and into a data file the program uses, so the replacements you do can be adjusted without re-compiling the code (this is the sort of thing that frequently seems to need a fair amount of "tweaking", since there's no one way of doing it that's obviously correct and the other ways are wrong).

\$\endgroup\$
  • \$\begingroup\$ Loading the data from a file is a good idea. About the map, since "á" and "à" would both turn into "a", do you think I should use a QMap<QString, QStringList>() instead or would that be a rather bad idea? \$\endgroup\$ – Davlog Jan 14 '16 at 17:36
  • 1
    \$\begingroup\$ @Davlog: I doubt it makes a whole lot of difference in either direction. \$\endgroup\$ – Jerry Coffin Jan 14 '16 at 18:31
0
\$\begingroup\$

I would eliminate lines and clean up the code by relying on regular expressions.

QString s = "áche über dir Ke$ha is worth $100";

// Performance: Eliminate characters you do not wish to have. 
s.remove(QRegularExpression("[" + QRegularExpression::escape("'!*,?|¡¿") + "]"));
qDebug().noquote() << "Before:\t" << s;

// Performance: Check for characters
if (s.contains(QRegularExpression("[" + QRegularExpression::escape("$/:ÀÁÄÙÛÜàáäçèéêëïñóöùûü") + "]")))
{
    // Special Characters 
    // Escape function is a safety measure in case you accidentally insert "^" in the square brackets.
    s.replace(QRegularExpression("[" + QRegularExpression::escape(":/") + "]"), "-");
    s.replace(QRegularExpression("[$]"), "s");

    // Upper Case
    s.replace(QRegularExpression("[ÁÀ]"),   "A");
    s.replace(QRegularExpression("[Ä]"),    "Ae");
    s.replace(QRegularExpression("[ÜÛÙ]"),  "U");

    // Lower Case
    s.replace(QRegularExpression("[áà]"),   "a");
    s.replace(QRegularExpression("[ä]"),    "ae");
    s.replace(QRegularExpression("[ç]"),    "c");
    s.replace(QRegularExpression("[ëêéè]"), "e");
    s.replace(QRegularExpression("[ï]"),    "i");
    s.replace(QRegularExpression("[ñ]"),    "n");
    s.replace(QRegularExpression("[óö]"),   "o");
    s.replace(QRegularExpression("[ûù]"),   "u");
    s.replace(QRegularExpression("[ü]"),    "ue");
}

qDebug().noquote() << " After:\t" << s;

Before:  áche über dir Ke$ha is worth $100
 After:  ache ueber dir Kesha is worth s100

Oops; found an error in your code. Lets just adjust this line then:

    s.replace(QRegularExpression("[$]([^0-9])"), "s\\1");

Before:  áche über dir Ke$ha is worth $100
 After:  ache ueber dir Kesha is worth $100
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.