正如@CraigTP在他的精彩回答中提到的，我也有类似的性能问题，使用exception来测试字符串是否是数值。因此，我最终分割字符串并使用java.lang.Character.isDigit()。

public static boolean isNumeric(String str)
{
    for (char c : str.toCharArray())
    {
        if (!Character.isDigit(c)) return false;
    }
    return true;
}

根据Javadoc, Character.isDigit(char)将正确识别非拉丁数字。在性能方面，我认为简单的N次比较(其中N是字符串中的字符数)在计算上比正则表达式匹配更有效。

更新:正如Jean-François Corbett在评论中指出的那样，上面的代码只验证正整数，这涵盖了我的大部分用例。下面是更新后的代码，它根据系统中使用的默认区域设置正确验证十进制数，假设十进制分隔符只在字符串中出现一次。

public static boolean isStringNumeric( String str )
{
    DecimalFormatSymbols currentLocaleSymbols = DecimalFormatSymbols.getInstance();
    char localeMinusSign = currentLocaleSymbols.getMinusSign();

    if ( !Character.isDigit( str.charAt( 0 ) ) && str.charAt( 0 ) != localeMinusSign ) return false;

    boolean isDecimalSeparatorFound = false;
    char localeDecimalSeparator = currentLocaleSymbols.getDecimalSeparator();

    for ( char c : str.substring( 1 ).toCharArray() )
    {
        if ( !Character.isDigit( c ) )
        {
            if ( c == localeDecimalSeparator && !isDecimalSeparatorFound )
            {
                isDecimalSeparatorFound = true;
                continue;
            }
            return false;
        }
    }
    return true;
}

谷歌的Guava库提供了一个很好的辅助方法:你可以像使用Integer一样使用它。parseInt，但如果字符串没有解析为有效整数，它将返回null而不是抛出异常。注意，它返回的是Integer，而不是int，所以你必须将它转换/自动装箱回int。

例子:

String s1 = "22";
String s2 = "22.2";
Integer oInt1 = Ints.tryParse(s1);
Integer oInt2 = Ints.tryParse(s2);

int i1 = -1;
if (oInt1 != null) {
    i1 = oInt1.intValue();
}
int i2 = -1;
if (oInt2 != null) {
    i2 = oInt2.intValue();
}

System.out.println(i1);  // prints 22
System.out.println(i2);  // prints -1

但是，在当前发行版(Guava r11)中，它仍然被标记为@Beta。

我还没有对它进行基准测试。查看源代码，有一些开销来自大量的完整性检查，但最终他们使用Character.digit(string.charAt(idx))，类似，但略有不同，从@Ibrahim上面的答案。在它们的实现中没有异常处理开销。

这是我对这个问题的回答。

一个方便的方法，你可以使用任何类型的解析器来解析任何字符串:isParsable(对象解析器，字符串str)。解析器可以是Class或对象。这也将允许你使用你写的自定义解析器，应该适用于任何场景，例如:

isParsable(Integer.class, "11");
isParsable(Double.class, "11.11");
Object dateFormater = new java.text.SimpleDateFormat("yyyy.MM.dd G 'at' HH:mm:ss z");
isParsable(dateFormater, "2001.07.04 AD at 12:08:56 PDT");

下面是我的代码和方法描述。

import java.lang.reflect.*;

/**
 * METHOD: isParsable<p><p>
 * 
 * This method will look through the methods of the specified <code>from</code> parameter
 * looking for a public method name starting with "parse" which has only one String
 * parameter.<p>
 * 
 * The <code>parser</code> parameter can be a class or an instantiated object, eg:
 * <code>Integer.class</code> or <code>new Integer(1)</code>. If you use a
 * <code>Class</code> type then only static methods are considered.<p>
 * 
 * When looping through potential methods, it first looks at the <code>Class</code> associated
 * with the <code>parser</code> parameter, then looks through the methods of the parent's class
 * followed by subsequent ancestors, using the first method that matches the criteria specified
 * above.<p>
 * 
 * This method will hide any normal parse exceptions, but throws any exceptions due to
 * programmatic errors, eg: NullPointerExceptions, etc. If you specify a <code>parser</code>
 * parameter which has no matching parse methods, a NoSuchMethodException will be thrown
 * embedded within a RuntimeException.<p><p>
 * 
 * Example:<br>
 * <code>isParsable(Boolean.class, "true");<br>
 * isParsable(Integer.class, "11");<br>
 * isParsable(Double.class, "11.11");<br>
 * Object dateFormater = new java.text.SimpleDateFormat("yyyy.MM.dd G 'at' HH:mm:ss z");<br>
 * isParsable(dateFormater, "2001.07.04 AD at 12:08:56 PDT");<br></code>
 * <p>
 * 
 * @param parser    The Class type or instantiated Object to find a parse method in.
 * @param str   The String you want to parse
 * 
 * @return true if a parse method was found and completed without exception
 * @throws java.lang.NoSuchMethodException If no such method is accessible 
 */
public static boolean isParsable(Object parser, String str) {
    Class theClass = (parser instanceof Class? (Class)parser: parser.getClass());
    boolean staticOnly = (parser == theClass), foundAtLeastOne = false;
    Method[] methods = theClass.getMethods();

    // Loop over methods
    for (int index = 0; index < methods.length; index++) {
        Method method = methods[index];

        // If method starts with parse, is public and has one String parameter.
        // If the parser parameter was a Class, then also ensure the method is static. 
        if(method.getName().startsWith("parse") &&
            (!staticOnly || Modifier.isStatic(method.getModifiers())) &&
            Modifier.isPublic(method.getModifiers()) &&
            method.getGenericParameterTypes().length == 1 &&
            method.getGenericParameterTypes()[0] == String.class)
        {
            try {
                foundAtLeastOne = true;
                method.invoke(parser, str);
                return true; // Successfully parsed without exception
            } catch (Exception exception) {
                // If invoke problem, try a different method
                /*if(!(exception instanceof IllegalArgumentException) &&
                   !(exception instanceof IllegalAccessException) &&
                   !(exception instanceof InvocationTargetException))
                        continue; // Look for other parse methods*/

                // Parse method refuses to parse, look for another different method
                continue; // Look for other parse methods
            }
        }
    }

    // No more accessible parse method could be found.
    if(foundAtLeastOne) return false;
    else throw new RuntimeException(new NoSuchMethodException());
}


/**
 * METHOD: willParse<p><p>
 * 
 * A convienence method which calls the isParseable method, but does not throw any exceptions
 * which could be thrown through programatic errors.<p>
 * 
 * Use of {@link #isParseable(Object, String) isParseable} is recommended for use so programatic
 * errors can be caught in development, unless the value of the <code>parser</code> parameter is
 * unpredictable, or normal programtic exceptions should be ignored.<p>
 * 
 * See {@link #isParseable(Object, String) isParseable} for full description of method
 * usability.<p>
 * 
 * @param parser    The Class type or instantiated Object to find a parse method in.
 * @param str   The String you want to parse
 * 
 * @return true if a parse method was found and completed without exception
 * @see #isParseable(Object, String) for full description of method usability 
 */
public static boolean willParse(Object parser, String str) {
    try {
        return isParsable(parser, str);
    } catch(Throwable exception) {
        return false;
    }
}

这是一个简单的检查示例:

public static boolean isNumericString(String input) {
    boolean result = false;

    if(input != null && input.length() > 0) {
        char[] charArray = input.toCharArray();

        for(char c : charArray) {
            if(c >= '0' && c <= '9') {
                // it is a digit
                result = true;
            } else {
                result = false;
                break;
            }
        }
    }

    return result;
}

public static boolean isNumeric(String str)
{
    return str.matches("-?\\d+(.\\d+)?");
}

CraigTP的正则表达式(如上所示)会产生一些误报。如。"23y4"将被算作一个数字，因为'。'匹配任何字符，而不是小数点。

此外，它将拒绝任何以“+”开头的数字

避免这两个小问题的替代方法是

public static boolean isNumeric(String str)
{
    return str.matches("[+-]?\\d*(\\.\\d+)?");
}

这里有两种可能有效的方法。(不使用异常)。 注意:Java默认是值传递，String的值是String对象数据的地址。 所以，当你在做

stringNumber = stringNumber.replaceAll(" ", "");

您已将输入值更改为没有空格。 如果你愿意，可以去掉这条线。

private boolean isValidStringNumber(String stringNumber)
{
    if(stringNumber.isEmpty())
    {
        return false;
    }

    stringNumber = stringNumber.replaceAll(" ", "");

    char [] charNumber = stringNumber.toCharArray();
    for(int i =0 ; i<charNumber.length ;i++)
    {
        if(!Character.isDigit(charNumber[i]))
        {
            return false;
        }
    }
    return true;
}

这里是另一个方法，以防你想允许浮动 据称，这种方法允许表单中的数字通过 1123123123123123年.123 我刚做好，我想还需要进一步测试以确保它能正常工作。

private boolean isValidStringTrueNumber(String stringNumber)
{
    if(stringNumber.isEmpty())
    {
        return false;
    }

    stringNumber = stringNumber.replaceAll(" ", "");
    int countOfDecimalPoint = 0;
    boolean decimalPointPassed = false;
    boolean commaFound = false;
    int countOfDigitsBeforeDecimalPoint = 0;
    int countOfDigitsAfterDecimalPoint =0 ;
    int commaCounter=0;
    int countOfDigitsBeforeFirstComma = 0;

    char [] charNumber = stringNumber.toCharArray();
    for(int i =0 ; i<charNumber.length ;i++)
    {
        if((commaCounter>3)||(commaCounter<0))
        {
            return false;
        }
        if(!Character.isDigit(charNumber[i]))//Char is not a digit.
        {
            if(charNumber[i]==',')
            {
                if(decimalPointPassed)
                {
                    return false;
                }
                commaFound = true;
                //check that next three chars are only digits.
                commaCounter +=3;
            }
            else if(charNumber[i]=='.')
            {
                decimalPointPassed = true;
                countOfDecimalPoint++;
            }
            else
            {
                return false;
            }
        }
        else //Char is a digit.
        {
            if ((commaCounter>=0)&&(commaFound))
            {
                if(!decimalPointPassed)
                {
                    commaCounter--;
                }
            }

            if(!commaFound)
            {
                countOfDigitsBeforeFirstComma++;
            }

            if(!decimalPointPassed)
            {
                countOfDigitsBeforeDecimalPoint++;
            }
            else
            {
                countOfDigitsAfterDecimalPoint++;
            }
        }
    }
    if((commaFound)&&(countOfDigitsBeforeFirstComma>3))
    {
        return false;
    }
    if(countOfDecimalPoint>1)
    {
        return false;
    }

    if((decimalPointPassed)&&((countOfDigitsBeforeDecimalPoint==0)||(countOfDigitsAfterDecimalPoint==0)))
    {
        return false;
    }
    return true;
}