保留带有贪婪令牌的 REGEX 定界符

Question

美好的一天，

我正在 Java 中编写方程式计算器并使用 REGEX 来识别值，包括我在其中一个提要中发现的科学记数法（并略微采用），如下所示：

[\d.]+(?:E-?\d+)?

我遇到的问题是我想保留分隔值。我怎样才能做到这一点？我在 regex101.com 上玩过它，但是，当我使用向前看和向后看时，它会抱怨贪婪的令牌。

我在 Whosebug 上找到了其他几个 REGEX，但找不到一个保留分隔符的。

提前致谢！

Answer 1

可能不是世界上最快的事情，但你可以像这样做一些：

/**
 * Holds onto and supplies the supplied split delimiter(s) to the split
 * array elements.<br><br>
 * <p>
 * This method creates a Regular Expression (RegEx) that is to be placed
 * within a String.split() method to acquire the desired array
 * content.<br><br>
 *
 * @param inputString       (String) The string to split.<br>
 *
 * @param delimiterPosition (Integer) A integer value of either 0, 1, or 2.
 *                          The specific value determines how the detected
 *                          delimiter types are placed within the array:<pre>
 *
 *      0       Delimiter as separate element:
 *              a;b;c;d = [a, ;, b, ;, c, ;, d]
 *              Core regex is: .split("((?&lt;=;)|(?=;))")
 *              Lookahead and Lookbehind used.
 *
 *      1       Delimiter at end of each element except last:
 *              a;b;c;d = [a;, b;, c;, d]
 *              Core regex is: .split("(?&lt;=;)")
 *              Lookahead used only.
 *
 *      2       Delimiter at beginning of each element except first:
 *              a;b;c;d = [a, ;b, ;c, ;d]
 *              Core regex is: .split("(?=;)")
 *              Lookbehind used only.</pre><br>
 *
 * If nothing is supplied then each character of the supplied input string
 * is split into the sting array.<br><br>
 *
 * If any supplied delimiters or delimiter characters happen to be RegEx
 * Meta Characters such as: ( ) [ ] { { \ ^ $ | ? * + . &lt; &gt; - = ! for
 * example then those delimiters must be Escaped with a Double Backslash
 * (ie: "\+" ) when supplied otherwise an exception will occur.<br>
 *
 * @param delimiters        (1D String Array or one to multiple comma
 *                          delimited String Entries) Any number of string
 *                          delimiters can be supplied as long as they are
 *                          separated with a comma (,).<br>
 *
 * @return (String) The Regular Expression (RegEx) to be used within a
 *         String.split() method.
 */
public static String[] SplitAndKeepDelimiters(String inputString, int delimiterPosition, String... delimiters) {
    if (delimiters.length < 1) {
        return inputString.split("");
    }

    // build regex...
    String regEx = "";
    for (int i = 0; i < delimiters.length; i++) {
        switch (delimiterPosition) {
            case 0:
                regEx += regEx.isEmpty() ? "((?<=" + delimiters[i] + ")|(?=" + delimiters[i] + "))"
                        : "|((?<=" + delimiters[i] + ")|(?=" + delimiters[i] + "))";
                break;
            case 1:
                regEx += regEx.isEmpty() ? "(?<=" + delimiters[i] + ")"
                        : "|(?<=" + delimiters[i] + ")";
                break;
            case 2:
                regEx += regEx.isEmpty() ? "(?=" + delimiters[i] + ")"
                        : "|(?=" + delimiters[i] + ")";
                break;
        }
    }
    return inputString.split(regEx);
}

上述方法将允许您拆分多个分隔符。

Answer 2

除了使用拆分，您还可以使用交替来获取匹配项，或者匹配所有未直接跟在第一个模式后面的字符。

[\d.]+(?:E-?\d+)?|(?:(?![\d.]+(?:E-?\d+)?).)+

模式匹配：

[\d.]+(?:E-?\d+)? 你的科学记数法模式
| 或
(?:非捕获组
- (?![\d.]+(?:E-?\d+)?).负前瞻，当科学记数法不在正右时匹配单个字符
)+关闭非捕获组，并重复1+次以匹配至少一个字符

Regex demo | Java demo

例如

String regex = "[\d.]+(?:E-?\d+)?|(?:(?![\d.]+(?:E-?\d+)?).)+";
String string = "cos(2123.324E3)*ln(e^x)+123.345E-6*sin(sin(sin(x)))";

Pattern pattern = Pattern.compile(regex, Pattern.MULTILINE);
Matcher matcher = pattern.matcher(string);

while (matcher.find()) {
    System.out.println(matcher.group(0));
}

输出

cos(
2123.324E3
)*ln(e^x)+
123.345E-6
*sin(sin(sin(x)))

保留带有贪婪令牌的 REGEX 定界符

Keep REGEX delimiter with greedy token

java

regex

scientific-notation

delimiter