views:

147

answers:

3

We currently have a crude mechanism to convert numbers to words (e.g. using a few static arrays) and based on the size of the number translating that into an english text. But we are running into issues for numbers which are huge.

10183 = Ten thousand one hundred eighty three
90 = Ninety
5888 = Five thousand eight hundred eighty eight

Is there an easy to use function in any of the math libraries which I can use for this purpose?

A: 

Because you cannot have a general algorithm for every locale, no. You have to implement your own algorithm for every locale that you are supporting.

** EDIT **

Just for the heck of it, I played around until I got this class. It cannot display Long.MIN_VALUE because of bit restrictions... but I presume it could be modified and change long value types to double for decimal or even bigger numbers It can display any numbers up to 66 digits and 26 decimals using a string representation of the number. You may add more ScaleUnit for more decimals...

/**
 * This class will convert numeric values into an english representation
 * 
 * For units, see : http://www.jimloy.com/math/billion.htm
 * 
 * @author [email protected]
 */
public class NumberToWords {

    static public class ScaleUnit {
        private int exponent;
        private String[] names;
        private ScaleUnit(int exponent, String...names) {
            this.exponent = exponent;
            this.names = names;
        }
        public int getExponent() {
            return exponent;
        }
        public String getName(int index) {
            return names[index];
        }
    }

    /**
     * See http://www.wordiq.com/definition/Names_of_large_numbers
     */
    static private ScaleUnit[] SCALE_UNITS = new ScaleUnit[] {
        new ScaleUnit(63, "vigintillion", "decilliard"),
        new ScaleUnit(60, "novemdecillion", "decillion"),
        new ScaleUnit(57, "octodecillion", "nonilliard"),
        new ScaleUnit(54, "septendecillion", "nonillion"),
        new ScaleUnit(51, "sexdecillion", "octilliard"),
        new ScaleUnit(48, "quindecillion", "octillion"),
        new ScaleUnit(45, "quattuordecillion", "septilliard"),
        new ScaleUnit(42, "tredecillion", "septillion"),
        new ScaleUnit(39, "duodecillion", "sextilliard"),
        new ScaleUnit(36, "undecillion", "sextillion"),
        new ScaleUnit(33, "decillion", "quintilliard"),
        new ScaleUnit(30, "nonillion", "quintillion"),
        new ScaleUnit(27, "octillion", "quadrilliard"),
        new ScaleUnit(24, "septillion", "quadrillion"),
        new ScaleUnit(21, "sextillion", "trilliard"),
        new ScaleUnit(18, "quintillion", "trillion"),
        new ScaleUnit(15, "quadrillion", "billiard"),
        new ScaleUnit(12, "trillion", "billion"),
        new ScaleUnit(9, "billion", "milliard"),
        new ScaleUnit(6, "million", "million"),
        new ScaleUnit(3, "thousand", "thousand"),
        new ScaleUnit(2, "hundred", "hundred"),
        //new ScaleUnit(1, "ten", "ten"),
        //new ScaleUnit(0, "one", "one"),
        new ScaleUnit(-1, "tenth", "tenth"),
        new ScaleUnit(-2, "hundredth", "hundredth"),
        new ScaleUnit(-3, "thousandth", "thousandth"),
        new ScaleUnit(-4, "ten-thousandth", "ten-thousandth"),
        new ScaleUnit(-5, "hundred-thousandth", "hundred-thousandth"),
        new ScaleUnit(-6, "millionth", "millionth"),
        new ScaleUnit(-7, "ten-millionth", "ten-millionth"),
        new ScaleUnit(-8, "hundred-millionth", "hundred-millionth"),
        new ScaleUnit(-9, "billionth", "milliardth"),
        new ScaleUnit(-10, "ten-billionth", "ten-milliardth"),
        new ScaleUnit(-11, "hundred-billionth", "hundred-milliardth"),
        new ScaleUnit(-12, "trillionth", "billionth"),
        new ScaleUnit(-13, "ten-trillionth", "ten-billionth"),
        new ScaleUnit(-14, "hundred-trillionth", "hundred-billionth"),
        new ScaleUnit(-15, "quadrillionth", "billiardth"),
        new ScaleUnit(-16, "ten-quadrillionth", "ten-billiardth"),
        new ScaleUnit(-17, "hundred-quadrillionth", "hundred-billiardth"),
        new ScaleUnit(-18, "quintillionth", "trillionth"),
        new ScaleUnit(-19, "ten-quintillionth", "ten-trillionth"),
        new ScaleUnit(-20, "hundred-quintillionth", "hundred-trillionth"),
        new ScaleUnit(-21, "sextillionth", "trilliardth"),
        new ScaleUnit(-22, "ten-sextillionth", "ten-trilliardth"),
        new ScaleUnit(-23, "hundred-sextillionth", "hundred-trilliardth"),
        new ScaleUnit(-24, "septillionth","quadrillionth"),
        new ScaleUnit(-25, "ten-septillionth","ten-quadrillionth"),
        new ScaleUnit(-26, "hundred-septillionth","hundred-quadrillionth"),
    };

    static public enum Scale {
        SHORT,
        LONG;

        public String getName(int exponent) {
            for (ScaleUnit unit : SCALE_UNITS) {
                if (unit.getExponent() == exponent) {
                    return unit.getName(this.ordinal());
                }
            }
            return ""; 
        }
    }

    /**
     * Change this scale to support American and modern British value (short scale)
     * or Traditional British value (long scale)
     */
    static public Scale SCALE = Scale.SHORT; 


    static abstract public class AbstractProcessor {

        static protected final String SEPARATOR = " ";
        static protected final int NO_VALUE = -1;

        protected List<Integer> getDigits(long value) {
            ArrayList<Integer> digits = new ArrayList<Integer>();
            if (value == 0) {
                digits.add(0);
            } else {
                while (value > 0) {
                    digits.add(0, (int) value % 10);
                    value /= 10;
                }
            }
            return digits;
        }

        public String getName(long value) {
            return getName(Long.toString(value));
        }

        public String getName(double value) {
            return getName(Double.toString(value));
        }

        abstract public String getName(String value);
    }

    static public class UnitProcessor extends AbstractProcessor {

        static private final String[] TOKENS = new String[] {
            "one", "two", "three", "four", "five", "six", "seven", "eight", "nine",
            "ten", "eleven", "twelve", "thirteen", "fourteen", "fifteen", "sixteen", "seventeen", "eighteen", "nineteen"
        };

        @Override
        public String getName(String value) {
            StringBuilder buffer = new StringBuilder();

            int offset = NO_VALUE;
            int number;
            if (value.length() > 3) {
                number = Integer.valueOf(value.substring(value.length() - 3), 10);
            } else {
                number = Integer.valueOf(value, 10);
            }
            number %= 100;
            if (number < 10) {
                offset = (number % 10) - 1;
                //number /= 10;
            } else if (number < 20) {
                offset = (number % 20) - 1;
                //number /= 100;
            }

            if (offset != NO_VALUE && offset < TOKENS.length) {
                buffer.append(TOKENS[offset]);
            }

            return buffer.toString();
        }

    }

    static public class TensProcessor extends AbstractProcessor {

        static private final String[] TOKENS = new String[] {
            "twenty", "thirty", "fourty", "fifty", "sixty", "seventy", "eighty", "ninety"
        };

        static private final String UNION_SEPARATOR = "-";

        private UnitProcessor unitProcessor = new UnitProcessor();

        @Override
        public String getName(String value) {
            StringBuilder buffer = new StringBuilder();
            boolean tensFound = false;

            int number;
            if (value.length() > 3) {
                number = Integer.valueOf(value.substring(value.length() - 3), 10);
            } else {
                number = Integer.valueOf(value, 10);
            }
            number %= 100;   // keep only two digits
            if (number >= 20) {
                buffer.append(TOKENS[(number / 10) - 2]);
                number %= 10;
                tensFound = true;
            } else {
                number %= 20;
            }

            if (number != 0) {
                if (tensFound) {
                    buffer.append(UNION_SEPARATOR);
                }
                buffer.append(unitProcessor.getName(number));
            }

            return buffer.toString();
        }
    }

    static public class HundredProcessor extends AbstractProcessor {

        private int EXPONENT = 2;

        private UnitProcessor unitProcessor = new UnitProcessor();
        private TensProcessor tensProcessor = new TensProcessor();

        @Override
        public String getName(String value) {
            StringBuilder buffer = new StringBuilder();

            int number;
            if (value.isEmpty()) {
                number = 0;
            } else if (value.length() > 4) {
                number = Integer.valueOf(value.substring(value.length() - 4), 10);
            } else {
                number = Integer.valueOf(value, 10);
            }
            number %= 1000;  // keep at least three digits

            if (number >= 100) {
                buffer.append(unitProcessor.getName(number / 100));
                buffer.append(SEPARATOR);
                buffer.append(SCALE.getName(EXPONENT));
            }

            String tensName = tensProcessor.getName(number % 100);

            if (!tensName.isEmpty() && (number >= 100)) {
                buffer.append(SEPARATOR);
            }
            buffer.append(tensName);

            return buffer.toString();
        }
    }

    static public class CompositeBigProcessor extends AbstractProcessor {

        private HundredProcessor hundredProcessor = new HundredProcessor();
        private AbstractProcessor lowProcessor;
        private int exponent;

        public CompositeBigProcessor(int exponent) {
            if (exponent <= 3) {
                lowProcessor = hundredProcessor;
            } else {
                lowProcessor = new CompositeBigProcessor(exponent - 3);
            }
            this.exponent = exponent;
        }

        public String getToken() {
            return SCALE.getName(getPartDivider());
        }

        protected AbstractProcessor getHighProcessor() {
            return hundredProcessor;
        }

        protected AbstractProcessor getLowProcessor() {
            return lowProcessor;
        }

        public int getPartDivider() {
            return exponent;
        }

        @Override
        public String getName(String value) {
            StringBuilder buffer = new StringBuilder();

            String high, low;
            if (value.length() < getPartDivider()) {
                high = "";
                low = value;
            } else {
                int index = value.length() - getPartDivider();
                high = value.substring(0, index);
                low = value.substring(index);
            }

            String highName = getHighProcessor().getName(high);
            String lowName = getLowProcessor().getName(low);

            if (!highName.isEmpty()) {
                buffer.append(highName);
                buffer.append(SEPARATOR);
                buffer.append(getToken());

                if (!lowName.isEmpty()) {
                    buffer.append(SEPARATOR);
                }
            }

            if (!lowName.isEmpty()) {
                buffer.append(lowName);
            }

            return buffer.toString();
        }
    }

    static public class DefaultProcessor extends AbstractProcessor {

        static private String MINUS = "minus";
        static private String UNION_AND = "and";

        static private String ZERO_TOKEN = "zero";

        private AbstractProcessor processor = new CompositeBigProcessor(63);

        @Override
        public String getName(String value) {
            boolean negative = false;
            if (value.startsWith("-")) {
                negative = true;
                value = value.substring(1);
            }

            int decimals = value.indexOf(".");
            String decimalValue = null;
            if (0 <= decimals) {
                decimalValue = value.substring(decimals + 1);
                value = value.substring(0, decimals);
            }

            String name = processor.getName(value);

            if (name.isEmpty()) {
                name = ZERO_TOKEN;
            } else if (negative) {
                name = MINUS.concat(SEPARATOR).concat(name); 
            }

            if (!(null == decimalValue || decimalValue.isEmpty())) {
                name = name.concat(SEPARATOR).concat(UNION_AND).concat(SEPARATOR)
                    .concat(processor.getName(decimalValue))
                    .concat(SEPARATOR).concat(SCALE.getName(-decimalValue.length()));
            }

            return name;
        }

    }

    static public AbstractProcessor processor;


    public static void main(String...args) {

        processor = new DefaultProcessor();

        long[] values = new long[] {
            0,
            4,
            10,
            12,
            100,
            108,
            299,
            1000,
            1003,
            2040,
            45213,
            100000,
            100005,
            100010,
            202020,
            202022,
            999999,
            1000000,
            1000001,
            10000000,
            10000007,
            99999999,
            Long.MAX_VALUE,
            Long.MIN_VALUE
        };

        String[] strValues = new String[] {
            "0001.2",
            "3.141592"
        };

        for (long val : values) {
            System.out.println(val + " = " + processor.getName(val) );
        }

        for (String strVal : strValues) {
            System.out.println(strVal + " = " + processor.getName(strVal) );
        }

        // generate a very big number...
        StringBuilder bigNumber = new StringBuilder();
        for (int d=0; d<66; d++) {
            bigNumber.append( (char) ((Math.random() * 10) + '0'));
        }
        bigNumber.append(".");
        for (int d=0; d<26; d++) {
            bigNumber.append( (char) ((Math.random() * 10) + '0'));
        }

        System.out.println(bigNumber.toString() + " = " + processor.getName(bigNumber.toString()));

    }

}

and a sample output (for the random big number generator)

0 = zero
4 = four
10 = ten
12 = twelve
100 = one hundred
108 = one hundred eight
299 = two hundred ninety-nine
1000 = one thousand
1003 = one thousand three
2040 = two thousand fourty
45213 = fourty-five thousand two hundred thirteen
100000 = one hundred thousand
100005 = one hundred thousand five
100010 = one hundred thousand ten
202020 = two hundred two thousand twenty
202022 = two hundred two thousand twenty-two
999999 = nine hundred ninety-nine thousand nine hundred ninety-nine
1000000 = one million
1000001 = one million one
10000000 = ten million
10000007 = ten million seven
99999999 = ninety-nine million nine hundred ninety-nine thousand nine hundred ninety-nine
9223372036854775807 = nine quintillion two hundred twenty-three quadrillion three hundred seventy-two trillion thirty-six billion eight hundred fifty-four million seven hundred seventy-five thousand eight hundred seven
-9223372036854775808 = minus nine quintillion two hundred twenty-three quadrillion three hundred seventy-two trillion thirty-six billion eight hundred fifty-four million seven hundred seventy-five thousand eight hundred eight
0001.2 = one and two tenth
3.141592 = three and one hundred fourty-one thousand five hundred ninety-two millionth
694780458103427072928672912656674465845126458162617425283733729646.85695031739734695391404376 = six hundred ninety-four vigintillion seven hundred eighty novemdecillion four hundred fifty-eight octodecillion one hundred three septendecillion four hundred twenty-seven sexdecillion seventy-two quindecillion nine hundred twenty-eight quattuordecillion six hundred seventy-two tredecillion nine hundred twelve duodecillion six hundred fifty-six undecillion six hundred seventy-four decillion four hundred sixty-five nonillion eight hundred fourty-five octillion one hundred twenty-six septillion four hundred fifty-eight sextillion one hundred sixty-two quintillion six hundred seventeen quadrillion four hundred twenty-five trillion two hundred eighty-three billion seven hundred thirty-three million seven hundred twenty-nine thousand six hundred fourty-six and eighty-five septillion six hundred ninety-five sextillion thirty-one quintillion seven hundred thirty-nine quadrillion seven hundred thirty-four trillion six hundred ninety-five billion three hundred ninety-one million four hundred four thousand three hundred seventy-six hundred-septillionth
Yanick Rochon
just need this for english alone
Jason
@Jason, I was answering the question "Is there an easy to use function in any of the math libraries which I can use for this purpose?"
Yanick Rochon
+1  A: 

Here is the code, I don't think there is any method in SE.

org.life.java
+3  A: 

ICU4J contains nice number-spellout support. The files with the "rules" can be easily edited, and it's no problem to add other languages (we did it e.g. for Polish and Russian).

Landei
ICU4J's data comes from CLDR http://unicode.org/cldr/ and already supports Polish and Russian. http://unicode.org/repos/cldr/trunk/common/rbnf/ Did you file a ticket if you found issues with the existing rules?
Steven R. Loomis
No, but we did this some years ago, and if I remember right, we needed some additional things. But good to know!
Landei
RBNF has had major overhauls a year or so ago, both in CLDR and ICU. Check it out.
Steven R. Loomis