Splits a String by Character type as returned by java.lang.Character.getType(char). - Java java.lang

Java examples for java.lang:String Split

  1. HOME
  2. Java
  3. java.lang
  4. String Split

Description

Splits a String by Character type as returned by java.lang.Character.getType(char).

Demo Code

/*//from w  w  w .  j  av a 2 s  . co  m
 * Copyright 2013 Guidewire Software, Inc.
 */
/**
 * This class is based, in part, on org.apache.commons.lang.StringUtils and is intended
 * to break the dependency on that project.
 *
 * @author <a href="http://jakarta.apache.org/turbine/">Apache Jakarta Turbine</a>
 * @author <a href="mailto:jon@latchkey.com">Jon S. Stevens</a>
 * @author Daniel L. Rall
 * @author <a href="mailto:gcoladonato@yahoo.com">Greg Coladonato</a>
 * @author <a href="mailto:ed@apache.org">Ed Korthof</a>
 * @author <a href="mailto:rand_mcneely@yahoo.com">Rand McNeely</a>
 * @author Stephen Colebourne
 * @author <a href="mailto:fredrik@westermarck.com">Fredrik Westermarck</a>
 * @author Holger Krauth
 * @author <a href="mailto:alex@purpletech.com">Alexander Day Chaffee</a>
 * @author <a href="mailto:hps@intermeta.de">Henning P. Schmiedehausen</a>
 * @author Arun Mammen Thomas
 * @author Gary Gregory
 * @author Phil Steitz
 * @author Al Chou
 * @author Michael Davey
 * @author Reuben Sivan
 * @author Chris Hyzer
 *  Johnson

 */
//package com.java2s;

import java.util.ArrayList;

import java.util.List;

public class Main {
    public static void main(String[] argv) {
        String str = "java2s.com";
        System.out.println(java.util.Arrays
                .toString(splitByCharacterTypeCamelCase(str)));
    }

    private static final String[] EMPTY_STRING_ARRAY = new String[0];

    /**
     * <p>Splits a String by Character type as returned by
     * <code>java.lang.Character.getType(char)</code>. Groups of contiguous
     * characters of the same type are returned as complete tokens, with the
     * following exception: the character of type
     * <code>Character.UPPERCASE_LETTER</code>, if any, immediately
     * preceding a token of type <code>Character.LOWERCASE_LETTER</code>
     * will belong to the following token rather than to the preceding, if any,
     * <code>Character.UPPERCASE_LETTER</code> token.
     * <pre>
     * splitByCharacterTypeCamelCase(null)         = null
     * splitByCharacterTypeCamelCase("")           = []
     * splitByCharacterTypeCamelCase("ab de fg")   = ["ab", " ", "de", " ", "fg"]
     * splitByCharacterTypeCamelCase("ab   de fg") = ["ab", "   ", "de", " ", "fg"]
     * splitByCharacterTypeCamelCase("ab:cd:ef")   = ["ab", ":", "cd", ":", "ef"]
     * splitByCharacterTypeCamelCase("number5")    = ["number", "5"]
     * splitByCharacterTypeCamelCase("fooBar")     = ["foo", "Bar"]
     * splitByCharacterTypeCamelCase("foo200Bar")  = ["foo", "200", "Bar"]
     * splitByCharacterTypeCamelCase("ASFRules")   = ["ASF", "Rules"]
     * </pre>
     * @param str the String to split, may be <code>null</code>
     * @return an array of parsed Strings, <code>null</code> if null String input
     * @since 2.4
     */
    public static String[] splitByCharacterTypeCamelCase(String str) {
        return splitByCharacterType(str, true);
    }

    /**
     * <p>Splits a String by Character type as returned by
     * <code>java.lang.Character.getType(char)</code>. Groups of contiguous
     * characters of the same type are returned as complete tokens.
     * <pre>
     * splitByCharacterType(null)         = null
     * splitByCharacterType("")           = []
     * splitByCharacterType("ab de fg")   = ["ab", " ", "de", " ", "fg"]
     * splitByCharacterType("ab   de fg") = ["ab", "   ", "de", " ", "fg"]
     * splitByCharacterType("ab:cd:ef")   = ["ab", ":", "cd", ":", "ef"]
     * splitByCharacterType("number5")    = ["number", "5"]
     * splitByCharacterType("fooBar")     = ["foo", "B", "ar"]
     * splitByCharacterType("foo200Bar")  = ["foo", "200", "B", "ar"]
     * splitByCharacterType("ASFRules")   = ["ASFR", "ules"]
     * </pre>
     * @param str the String to split, may be <code>null</code>
     * @return an array of parsed Strings, <code>null</code> if null String input
     * @since 2.4
     */
    public static String[] splitByCharacterType(String str) {
        return splitByCharacterType(str, false);
    }

    /**
     * <p>Splits a String by Character type as returned by
     * <code>java.lang.Character.getType(char)</code>. Groups of contiguous
     * characters of the same type are returned as complete tokens, with the
     * following exception: if <code>camelCase</code> is <code>true</code>,
     * the character of type <code>Character.UPPERCASE_LETTER</code>, if any,
     * immediately preceding a token of type <code>Character.LOWERCASE_LETTER</code>
     * will belong to the following token rather than to the preceding, if any,
     * <code>Character.UPPERCASE_LETTER</code> token.
     * @param str the String to split, may be <code>null</code>
     * @param camelCase whether to use so-called "camel-case" for letter types
     * @return an array of parsed Strings, <code>null</code> if null String input
     * @since 2.4
     */
    private static String[] splitByCharacterType(String str,
            boolean camelCase) {
        if (str == null) {
            return null;
        }
        if (str.length() == 0) {
            return EMPTY_STRING_ARRAY;
        }
        char[] c = str.toCharArray();
        List list = new ArrayList();
        int tokenStart = 0;
        int currentType = Character.getType(c[tokenStart]);
        for (int pos = tokenStart + 1; pos < c.length; pos++) {
            int type = Character.getType(c[pos]);
            if (type == currentType) {
                continue;
            }
            if (camelCase && type == Character.LOWERCASE_LETTER
                    && currentType == Character.UPPERCASE_LETTER) {
                int newTokenStart = pos - 1;
                if (newTokenStart != tokenStart) {
                    list.add(new String(c, tokenStart, newTokenStart
                            - tokenStart));
                    tokenStart = newTokenStart;
                }
            } else {
                list.add(new String(c, tokenStart, pos - tokenStart));
                tokenStart = pos;
            }
            currentType = type;
        }
        list.add(new String(c, tokenStart, c.length - tokenStart));
        return (String[]) list.toArray(new String[list.size()]);
    }

    /**
     * Gets a String's length or <code>0</code> if the String is <code>null</code>.
     *
     * @param str
     *            a String or <code>null</code>
     * @return String length or <code>0</code> if the String is <code>null</code>.
     * @since 2.4
     */
    public static int length(String str) {
        return str == null ? 0 : str.length();
    }
}

Related Tutorials