Ignoring diacritic characters when comparing words with special characters (é, è, ...)

Question

I have a list with some Belgian cities with diacritic characters: (Liège, Quiévrain, Franière, etc.) and I would like to transform these special characters to compare with a list containing the same names in upper case, but without the diacritical marks (LIEGE, QUIEVRAIN, FRANIERE)

What i first tried to do was to use the upper case:

LIEGE.contentEqual(Liège.toUpperCase()) but that doesn't fit because the Upper case of Liège is LIÈGE and not LIEGE.

I have some complicated ideas like replacing each character, but that sounds stupid and a long process.

Any ideas on how to do this in a smart way?

Answer 1

As of Java 6, you can use java.text.Normalizer:

public String unaccent(String s) {
    String normalized = Normalizer.normalize(s, Normalizer.Form.NFD);
    return normalized.replaceAll("[^\\p{ASCII}]", "");
}

Note that in Java 5 there is also a sun.text.Normalizer, but its use is strongly discouraged since it's part of Sun's proprietary API and has been removed in Java 6.

Answer 2

Check out this method in Java

private static final String PLAIN_ASCII = "AaEeIiOoUu" // grave
            + "AaEeIiOoUuYy" // acute
            + "AaEeIiOoUuYy" // circumflex
            + "AaOoNn" // tilde
            + "AaEeIiOoUuYy" // umlaut
            + "Aa" // ring
            + "Cc" // cedilla
            + "OoUu" // double acute
    ;

    private static final String UNICODE = "\u00C0\u00E0\u00C8\u00E8\u00CC\u00EC\u00D2\u00F2\u00D9\u00F9"
            + "\u00C1\u00E1\u00C9\u00E9\u00CD\u00ED\u00D3\u00F3\u00DA\u00FA\u00DD\u00FD"
            + "\u00C2\u00E2\u00CA\u00EA\u00CE\u00EE\u00D4\u00F4\u00DB\u00FB\u0176\u0177"
            + "\u00C3\u00E3\u00D5\u00F5\u00D1\u00F1"
            + "\u00C4\u00E4\u00CB\u00EB\u00CF\u00EF\u00D6\u00F6\u00DC\u00FC\u0178\u00FF"
            + "\u00C5\u00E5" + "\u00C7\u00E7" + "\u0150\u0151\u0170\u0171";

    /**
     * remove accented from a string and replace with ascii equivalent
     */
    public static String removeAccents(String s) {
        if (s == null)
            return null;
        StringBuilder sb = new StringBuilder(s.length());
        int n = s.length();
        int pos = -1;
        char c;
        boolean found = false;
        for (int i = 0; i < n; i++) {
            pos = -1;
            c = s.charAt(i);
            pos = (c <= 126) ? -1 : UNICODE.indexOf(c);
            if (pos > -1) {
                found = true;
                sb.append(PLAIN_ASCII.charAt(pos));
            } else {
                sb.append(c);
            }
        }
        if (!found) {
            return s;
        } else {
            return sb.toString();
        }
    }

Answer 3

This is the simplest solution I've found so far and it works perfectly in our applications.

Normalizer.normalize(string, Normalizer.Form.NFD).replaceAll("\\p{InCombiningDiacriticalMarks}+", ""); 

But I don't know if the Normalizer is available on the Android platform.

Answer 4

If you still need that for Android API 8 or lower (Android 2.2, Java 1.5) where you don't have Normalizer class, here's my code, I think better to modify than Pentium10 answer:

public class StringAccentRemover {

    @SuppressWarnings("serial")
    private static final HashMap<Character, Character> accents  = new HashMap<Character, Character>(){
        {
            put('Ą', 'A');
            put('Ę', 'E');
            put('Ć', 'C');
            put('Ł', 'L');
            put('Ń', 'N');
            put('Ó', 'O');
            put('Ś', 'S');
            put('Ż', 'Z');
            put('Ź', 'Z');

            put('ą', 'a');
            put('ę', 'e');
            put('ć', 'c');
            put('ł', 'l');
            put('ń', 'n');
            put('ó', 'o');
            put('ś', 's');
            put('ż', 'z');
            put('ź', 'z');
        }
    };
    /**
     * remove accented from a string and replace with ascii equivalent
     */
    public static String removeAccents(String s) {
        char[] result = s.toCharArray();
        for(int i=0; i<result.length; i++) {
            Character replacement = accents.get(result[i]);
            if (replacement!=null) result[i] = replacement;
        }
        return new String(result);
    }

}

Answer 5

I don't know if it is avaible on Android but on the JVM, you should not reimplement it in your project and reuse already existing code: just use org.apache.commons.lang3.StringUtils#stripAccents

Answer 6

The Collator class is a good way to do it (see corresponding javadoc). Here is a unit test that shows how to use it :

import static org.junit.Assert.assertEquals;

import java.text.Collator;
import java.util.Locale;

import org.junit.Test;

public class CollatorTest {
    @Test public void liege() throws Exception {
        Collator compareOperator = Collator.getInstance(Locale.FRENCH);
        compareOperator.setStrength(Collator.PRIMARY);

        assertEquals(0, compareOperator.compare("Liege", "Liege")); // no accent
        assertEquals(0, compareOperator.compare("Liège", "Liege")); // with accent
        assertEquals(0, compareOperator.compare("LIEGE", "Liege")); // case insensitive
        assertEquals(0, compareOperator.compare("LIEGE", "Liège")); // case insensitive with accent

        assertEquals(1, compareOperator.compare("Liege", "Bruxelles"));
        assertEquals(-1, compareOperator.compare("Bruxelles", "Liege"));
    }
}

---

EDIT : sorry to see my answer did not meet your needs ; maybe it's beause I've presented it as unit test ? Is this ok for you ? I personnaly find it better because it's short and it uses the SDK (no need for String replacement)

Collator compareOperator = Collator.getInstance(Locale.FRENCH);
compareOperator.setStrength(Collator.PRIMARY);
if (compareOperator.compare("Liège", "Liege") == 0) {
    // if we are here, then it's the "same" String
}

hope this helps

Answer 7

For those looking for a clean java solution, use apache commons:

StringUtils.stripAccents("Liège").toUpperCase();

this will return

LIEGE

Answer 8

Since class Normalizer is not supported in Froyo or previous Android versions, I have combined this and this (which I both voted up), and optimized it, obtaining a couple of helper methods. Method unaccentify simply converts diacritic chars to plain chars, while method slugify generates a slug for the input string. Hope it can be useful to someone. Here is the source code:

import java.util.Arrays;
import java.util.Locale;  
import java.util.regex.Pattern;  

public class SlugFroyo {
    private static final Pattern STRANGE = Pattern.compile("[^a-zA-Z0-9-]");
    private static final Pattern WHITESPACE = Pattern.compile("[\\s]");

    private static final String DIACRITIC_CHARS = "\u00C0\u00E0\u00C8\u00E8\u00CC\u00EC\u00D2\u00F2\u00D9\u00F9"
            + "\u00C1\u00E1\u00C9\u00E9\u00CD\u00ED\u00D3\u00F3\u00DA\u00FA\u00DD\u00FD"
            + "\u00C2\u00E2\u00CA\u00EA\u00CE\u00EE\u00D4\u00F4\u00DB\u00FB\u0176\u0177"
            + "\u00C3\u00E3\u00D5\u00F5\u00D1\u00F1"
            + "\u00C4\u00E4\u00CB\u00EB\u00CF\u00EF\u00D6\u00F6\u00DC\u00FC\u0178\u00FF"
            + "\u00C5\u00E5" + "\u00C7\u00E7" + "\u0150\u0151\u0170\u0171";

    private static final String PLAIN_CHARS = "AaEeIiOoUu" // grave
            + "AaEeIiOoUuYy" // acute
            + "AaEeIiOoUuYy" // circumflex
            + "AaOoNn" // tilde
            + "AaEeIiOoUuYy" // umlaut
            + "Aa" // ring
            + "Cc" // cedilla
            + "OoUu"; // double acute

    private static char[] lookup = new char[0x180];

    static {
        Arrays.fill(lookup, (char) 0);
        for (int i = 0; i < DIACRITIC_CHARS.length(); i++)
            lookup[DIACRITIC_CHARS.charAt(i)] = PLAIN_CHARS.charAt(i);
    }

    public static String slugify(String s) {
        String nowhitespace = WHITESPACE.matcher(s).replaceAll("-");
        String unaccented = unaccentify(nowhitespace);
        String slug = STRANGE.matcher(unaccented).replaceAll("");
        return slug.toLowerCase(Locale.ENGLISH);
    }

    public static String unaccentify(String s) {
        StringBuilder sb = new StringBuilder(s);
        for (int i = 0; i < sb.length(); i++) {
            char c = sb.charAt(i);
            if (c > 126 && c < lookup.length) {
                char replacement = lookup[c];
                if (replacement > 0)
                    sb.setCharAt(i, replacement);
            }
        }
        return sb.toString();
    }
}