How to unescape HTML character entities in Java?

I tried Apache Commons StringEscapeUtils.unescapeHtml3() in my project, but wasn't satisfied with its performance. Turns out, it does a lot of unnecessary operations. For one, it allocates a StringWriter for every call, even if there's nothing to unescape in the string. I've rewritten that code differently, now it works much faster. Whoever finds this in google is welcome to use it.

Following code unescapes all HTML 3 symbols and numeric escapes (equivalent to Apache unescapeHtml3). You can just add more entries to the map if you need HTML 4.

package com.example;

import java.io.StringWriter;
import java.util.HashMap;

public class StringUtils {

    public static final String unescapeHtml3(final String input) {
        StringWriter writer = null;
        int len = input.length();
        int i = 1;
        int st = 0;
        while (true) {
            // look for '&'
            while (i < len && input.charAt(i-1) != '&')
                i++;
            if (i >= len)
                break;

            // found '&', look for ';'
            int j = i;
            while (j < len && j < i + MAX_ESCAPE + 1 && input.charAt(j) != ';')
                j++;
            if (j == len || j < i + MIN_ESCAPE || j == i + MAX_ESCAPE + 1) {
                i++;
                continue;
            }

            // found escape 
            if (input.charAt(i) == '#') {
                // numeric escape
                int k = i + 1;
                int radix = 10;

                final char firstChar = input.charAt(k);
                if (firstChar == 'x' || firstChar == 'X') {
                    k++;
                    radix = 16;
                }

                try {
                    int entityValue = Integer.parseInt(input.substring(k, j), radix);

                    if (writer == null) 
                        writer = new StringWriter(input.length());
                    writer.append(input.substring(st, i - 1));

                    if (entityValue > 0xFFFF) {
                        final char[] chrs = Character.toChars(entityValue);
                        writer.write(chrs[0]);
                        writer.write(chrs[1]);
                    } else {
                        writer.write(entityValue);
                    }

                } catch (NumberFormatException ex) { 
                    i++;
                    continue;
                }
            }
            else {
                // named escape
                CharSequence value = lookupMap.get(input.substring(i, j));
                if (value == null) {
                    i++;
                    continue;
                }

                if (writer == null) 
                    writer = new StringWriter(input.length());
                writer.append(input.substring(st, i - 1));

                writer.append(value);
            }

            // skip escape
            st = j + 1;
            i = st;
        }

        if (writer != null) {
            writer.append(input.substring(st, len));
            return writer.toString();
        }
        return input;
    }

    private static final String[][] ESCAPES = {
        {"\"",     "quot"}, // " - double-quote
        {"&",      "amp"}, // & - ampersand
        {"<",      "lt"}, // < - less-than
        {">",      "gt"}, // > - greater-than

        // Mapping to escape ISO-8859-1 characters to their named HTML 3.x equivalents.
        {"\u00A0", "nbsp"}, // non-breaking space
        {"\u00A1", "iexcl"}, // inverted exclamation mark
        {"\u00A2", "cent"}, // cent sign
        {"\u00A3", "pound"}, // pound sign
        {"\u00A4", "curren"}, // currency sign
        {"\u00A5", "yen"}, // yen sign = yuan sign
        {"\u00A6", "brvbar"}, // broken bar = broken vertical bar
        {"\u00A7", "sect"}, // section sign
        {"\u00A8", "uml"}, // diaeresis = spacing diaeresis
        {"\u00A9", "copy"}, // © - copyright sign
        {"\u00AA", "ordf"}, // feminine ordinal indicator
        {"\u00AB", "laquo"}, // left-pointing double angle quotation mark = left pointing guillemet
        {"\u00AC", "not"}, // not sign
        {"\u00AD", "shy"}, // soft hyphen = discretionary hyphen
        {"\u00AE", "reg"}, // ® - registered trademark sign
        {"\u00AF", "macr"}, // macron = spacing macron = overline = APL overbar
        {"\u00B0", "deg"}, // degree sign
        {"\u00B1", "plusmn"}, // plus-minus sign = plus-or-minus sign
        {"\u00B2", "sup2"}, // superscript two = superscript digit two = squared
        {"\u00B3", "sup3"}, // superscript three = superscript digit three = cubed
        {"\u00B4", "acute"}, // acute accent = spacing acute
        {"\u00B5", "micro"}, // micro sign
        {"\u00B6", "para"}, // pilcrow sign = paragraph sign
        {"\u00B7", "middot"}, // middle dot = Georgian comma = Greek middle dot
        {"\u00B8", "cedil"}, // cedilla = spacing cedilla
        {"\u00B9", "sup1"}, // superscript one = superscript digit one
        {"\u00BA", "ordm"}, // masculine ordinal indicator
        {"\u00BB", "raquo"}, // right-pointing double angle quotation mark = right pointing guillemet
        {"\u00BC", "frac14"}, // vulgar fraction one quarter = fraction one quarter
        {"\u00BD", "frac12"}, // vulgar fraction one half = fraction one half
        {"\u00BE", "frac34"}, // vulgar fraction three quarters = fraction three quarters
        {"\u00BF", "iquest"}, // inverted question mark = turned question mark
        {"\u00C0", "Agrave"}, // А - uppercase A, grave accent
        {"\u00C1", "Aacute"}, // Б - uppercase A, acute accent
        {"\u00C2", "Acirc"}, // В - uppercase A, circumflex accent
        {"\u00C3", "Atilde"}, // Г - uppercase A, tilde
        {"\u00C4", "Auml"}, // Д - uppercase A, umlaut
        {"\u00C5", "Aring"}, // Е - uppercase A, ring
        {"\u00C6", "AElig"}, // Ж - uppercase AE
        {"\u00C7", "Ccedil"}, // З - uppercase C, cedilla
        {"\u00C8", "Egrave"}, // И - uppercase E, grave accent
        {"\u00C9", "Eacute"}, // Й - uppercase E, acute accent
        {"\u00CA", "Ecirc"}, // К - uppercase E, circumflex accent
        {"\u00CB", "Euml"}, // Л - uppercase E, umlaut
        {"\u00CC", "Igrave"}, // М - uppercase I, grave accent
        {"\u00CD", "Iacute"}, // Н - uppercase I, acute accent
        {"\u00CE", "Icirc"}, // О - uppercase I, circumflex accent
        {"\u00CF", "Iuml"}, // П - uppercase I, umlaut
        {"\u00D0", "ETH"}, // Р - uppercase Eth, Icelandic
        {"\u00D1", "Ntilde"}, // С - uppercase N, tilde
        {"\u00D2", "Ograve"}, // Т - uppercase O, grave accent
        {"\u00D3", "Oacute"}, // У - uppercase O, acute accent
        {"\u00D4", "Ocirc"}, // Ф - uppercase O, circumflex accent
        {"\u00D5", "Otilde"}, // Х - uppercase O, tilde
        {"\u00D6", "Ouml"}, // Ц - uppercase O, umlaut
        {"\u00D7", "times"}, // multiplication sign
        {"\u00D8", "Oslash"}, // Ш - uppercase O, slash
        {"\u00D9", "Ugrave"}, // Щ - uppercase U, grave accent
        {"\u00DA", "Uacute"}, // Ъ - uppercase U, acute accent
        {"\u00DB", "Ucirc"}, // Ы - uppercase U, circumflex accent
        {"\u00DC", "Uuml"}, // Ь - uppercase U, umlaut
        {"\u00DD", "Yacute"}, // Э - uppercase Y, acute accent
        {"\u00DE", "THORN"}, // Ю - uppercase THORN, Icelandic
        {"\u00DF", "szlig"}, // Я - lowercase sharps, German
        {"\u00E0", "agrave"}, // а - lowercase a, grave accent
        {"\u00E1", "aacute"}, // б - lowercase a, acute accent
        {"\u00E2", "acirc"}, // в - lowercase a, circumflex accent
        {"\u00E3", "atilde"}, // г - lowercase a, tilde
        {"\u00E4", "auml"}, // д - lowercase a, umlaut
        {"\u00E5", "aring"}, // е - lowercase a, ring
        {"\u00E6", "aelig"}, // ж - lowercase ae
        {"\u00E7", "ccedil"}, // з - lowercase c, cedilla
        {"\u00E8", "egrave"}, // и - lowercase e, grave accent
        {"\u00E9", "eacute"}, // й - lowercase e, acute accent
        {"\u00EA", "ecirc"}, // к - lowercase e, circumflex accent
        {"\u00EB", "euml"}, // л - lowercase e, umlaut
        {"\u00EC", "igrave"}, // м - lowercase i, grave accent
        {"\u00ED", "iacute"}, // н - lowercase i, acute accent
        {"\u00EE", "icirc"}, // о - lowercase i, circumflex accent
        {"\u00EF", "iuml"}, // п - lowercase i, umlaut
        {"\u00F0", "eth"}, // р - lowercase eth, Icelandic
        {"\u00F1", "ntilde"}, // с - lowercase n, tilde
        {"\u00F2", "ograve"}, // т - lowercase o, grave accent
        {"\u00F3", "oacute"}, // у - lowercase o, acute accent
        {"\u00F4", "ocirc"}, // ф - lowercase o, circumflex accent
        {"\u00F5", "otilde"}, // х - lowercase o, tilde
        {"\u00F6", "ouml"}, // ц - lowercase o, umlaut
        {"\u00F7", "divide"}, // division sign
        {"\u00F8", "oslash"}, // ш - lowercase o, slash
        {"\u00F9", "ugrave"}, // щ - lowercase u, grave accent
        {"\u00FA", "uacute"}, // ъ - lowercase u, acute accent
        {"\u00FB", "ucirc"}, // ы - lowercase u, circumflex accent
        {"\u00FC", "uuml"}, // ь - lowercase u, umlaut
        {"\u00FD", "yacute"}, // э - lowercase y, acute accent
        {"\u00FE", "thorn"}, // ю - lowercase thorn, Icelandic
        {"\u00FF", "yuml"}, // я - lowercase y, umlaut
    };

    private static final int MIN_ESCAPE = 2;
    private static final int MAX_ESCAPE = 6;

    private static final HashMap<String, CharSequence> lookupMap;
    static {
        lookupMap = new HashMap<String, CharSequence>();
        for (final CharSequence[] seq : ESCAPES) 
            lookupMap.put(seq[1].toString(), seq[0]);
    }

}

I have used the Apache Commons StringEscapeUtils.unescapeHtml4() for this:

Unescapes a string containing entity escapes to a string containing the actual Unicode characters corresponding to the escapes. Supports HTML 4.0 entities.


The libraries mentioned in other answers would be fine solutions, but if you already happen to be digging through real-world html in your project, the Jsoup project has a lot more to offer than just managing "ampersand pound FFFF semicolon" things.

// textValue: <p>This is a&nbsp;sample. \"Granny\" Smith &#8211;.<\/p>\r\n
// becomes this: This is a sample. "Granny" Smith –.
// with one line of code:
// Jsoup.parse(textValue).getText(); // for older versions of Jsoup
Jsoup.parse(textValue).text();

// Another possibility may be the static unescapeEntities method:
boolean strictMode = true;
String unescapedString = org.jsoup.parser.Parser.unescapeEntities(textValue, strictMode);

And you also get the convenient API for extracting and manipulating data, using the best of DOM, CSS, and jquery-like methods. It's open source and MIT licence.