blade.kit.CharKit Maven / Gradle / Ivy
/**
* Copyright (c) 2015, biezhi 王爵 ([email protected])
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package blade.kit;
import java.io.UnsupportedEncodingException;
/**
* 字符相关的工具类。
*
* @author biezhi
* @since 1.0
*/
public abstract class CharKit {
private static final String CHAR_STRING = "\u0000\u0001\u0002\u0003\u0004\u0005\u0006\u0007"
+ "\b\t\n\u000b\f\r\u000e\u000f" + "\u0010\u0011\u0012\u0013\u0014\u0015\u0016\u0017"
+ "\u0018\u0019\u001a\u001b\u001c\u001d\u001e\u001f" + "\u0020\u0021\"\u0023\u0024\u0025\u0026\u0027"
+ "\u0028\u0029\u002a\u002b\u002c\u002d\u002e\u002f" + "\u0030\u0031\u0032\u0033\u0034\u0035\u0036\u0037"
+ "\u0038\u0039\u003a\u003b\u003c\u003d\u003e\u003f" + "\u0040\u0041\u0042\u0043\u0044\u0045\u0046\u0047"
+ "\u0048\u0049\u004a\u004b\u004c\u004d\u004e\u004f" + "\u0050\u0051\u0052\u0053\u0054\u0055\u0056\u0057"
+ "\u0058\u0059\u005a\u005b\\\u005d\u005e\u005f" + "\u0060\u0061\u0062\u0063\u0064\u0065\u0066\u0067"
+ "\u0068\u0069\u006a\u006b\u006c\u006d\u006e\u006f" + "\u0070\u0071\u0072\u0073\u0074\u0075\u0076\u0077"
+ "\u0078\u0079\u007a\u007b\u007c\u007d\u007e\u007f";
private static final String[] CHAR_STRING_ARRAY = new String[128];
private static final Character[] CHAR_ARRAY = new Character[128];
static {
for (int i = 127; i >= 0; i--) {
CHAR_STRING_ARRAY[i] = CHAR_STRING.substring(i, i + 1);
CHAR_ARRAY[i] = new Character((char) i);
}
}
/**
* 将字符转成ASCII码
*
* @param ch 字符
* @return ASCII码
*/
public static int toAscii(char ch) {
if (ch <= 0xFF) {
return ch;
}
return 0x3F;
}
/**
* 是否为“空白字符”
*
* @param ch 字符
* @return 如果是“空白字符”则返回true
*/
public static boolean isWhitespace(char ch) {
return ch <= ' ';
}
public static Character toCharacterObject(char ch) {
if (ch < CHAR_ARRAY.length) {
return CHAR_ARRAY[ch];
}
return new Character(ch);
}
public static Character toCharacterObject(String str) {
if (StringKit.isEmpty(str)) {
return null;
}
return toCharacterObject(str.charAt(0));
}
public static char toChar(byte b) {
return (char) (b & 0xFF);
}
public static char toChar(Character ch) {
if (ch == null) {
throw new IllegalArgumentException("The Character must not be null");
}
return ch.charValue();
}
public static char toChar(Character ch, char defaultValue) {
if (ch == null) {
return defaultValue;
}
return ch.charValue();
}
// -----------------------------------------------------------------------
public static char toChar(String str) {
if (StringKit.isEmpty(str)) {
throw new IllegalArgumentException("The String must not be empty");
}
return str.charAt(0);
}
public static char toChar(String str, char defaultValue) {
if (StringKit.isEmpty(str)) {
return defaultValue;
}
return str.charAt(0);
}
public static int toIntValue(char ch) {
if (!isDigit(ch)) {
throw new IllegalArgumentException("The character " + ch + " is not in the range '0' - '9'");
}
return ch - 48;
}
public static int toIntValue(char ch, int defaultValue) {
if (!isDigit(ch)) {
return defaultValue;
}
return ch - 48;
}
public static int toIntValue(Character ch) {
if (ch == null) {
throw new IllegalArgumentException("The character must not be null");
}
return toIntValue(ch.charValue());
}
public static int toIntValue(Character ch, int defaultValue) {
if (ch == null) {
return defaultValue;
}
return toIntValue(ch.charValue(), defaultValue);
}
public static String toString(char ch) {
if (ch < 128) {
return CHAR_STRING_ARRAY[ch];
}
return new String(new char[] { ch });
}
public static String toString(Character ch) {
if (ch == null) {
return null;
}
return toString(ch.charValue());
}
public static String unicodeEscaped(char ch) {
if (ch < 0x10) {
return "\\u000" + Integer.toHexString(ch);
} else if (ch < 0x100) {
return "\\u00" + Integer.toHexString(ch);
} else if (ch < 0x1000) {
return "\\u0" + Integer.toHexString(ch);
}
return "\\u" + Integer.toHexString(ch);
}
public static String unicodeEscaped(Character ch) {
if (ch == null) {
return null;
}
return unicodeEscaped(ch.charValue());
}
public static byte[] toSimpleByteArray(char[] carr) {
if (carr == null) {
return null;
}
byte[] barr = new byte[carr.length];
for (int i = 0; i < carr.length; i++) {
barr[i] = (byte) carr[i];
}
return barr;
}
/**
* Converts char sequence into byte array.
*
* @see #toSimpleByteArray(char[])
*/
public static byte[] toSimpleByteArray(CharSequence charSequence) {
if (charSequence == null) {
return null;
}
byte[] barr = new byte[charSequence.length()];
for (int i = 0; i < barr.length; i++) {
barr[i] = (byte) charSequence.charAt(i);
}
return barr;
}
public static char[] toSimpleCharArray(byte[] barr) {
if (barr == null) {
return null;
}
char[] carr = new char[barr.length];
for (int i = 0; i < barr.length; i++) {
carr[i] = (char) (barr[i] & 0xFF);
}
return carr;
}
public static byte[] toAsciiByteArray(char[] carr) {
if (carr == null) {
return null;
}
byte[] barr = new byte[carr.length];
for (int i = 0; i < carr.length; i++) {
barr[i] = (byte) ((int) (carr[i] <= 0xFF ? carr[i] : 0x3F));
}
return barr;
}
public static byte[] toAsciiByteArray(CharSequence charSequence) {
if (charSequence == null) {
return null;
}
byte[] barr = new byte[charSequence.length()];
for (int i = 0; i < barr.length; i++) {
char c = charSequence.charAt(i);
barr[i] = (byte) ((int) (c <= 0xFF ? c : 0x3F));
}
return barr;
}
public static byte[] toRawByteArray(char[] carr) {
if (carr == null) {
return null;
}
byte[] barr = new byte[carr.length << 1];
for (int i = 0, bpos = 0; i < carr.length; i++) {
char c = carr[i];
barr[bpos++] = (byte) ((c & 0xFF00) >> 8);
barr[bpos++] = (byte) (c & 0x00FF);
}
return barr;
}
public static char[] toRawCharArray(byte[] barr) {
if (barr == null) {
return null;
}
int carrLen = barr.length >> 1;
if (carrLen << 1 < barr.length) {
carrLen++;
}
char[] carr = new char[carrLen];
int i = 0, j = 0;
while (i < barr.length) {
char c = (char) (barr[i] << 8);
i++;
if (i != barr.length) {
c += barr[i] & 0xFF;
i++;
}
carr[j++] = c;
}
return carr;
}
public static byte[] toByteArray(char[] carr) throws UnsupportedEncodingException {
if (carr == null) {
return null;
}
return new String(carr).getBytes("UTF-8");
}
public static byte[] toByteArray(char[] carr, String charset) throws UnsupportedEncodingException {
if (carr == null) {
return null;
}
return new String(carr).getBytes(charset);
}
public static char[] toCharArray(byte[] barr) throws UnsupportedEncodingException {
if (barr == null) {
return null;
}
return new String(barr, "UTF-8").toCharArray();
}
public static char[] toCharArray(byte[] barr, String charset) throws UnsupportedEncodingException {
if (barr == null) {
return null;
}
return new String(barr, charset).toCharArray();
}
public static boolean equalsOne(char c, char[] match) {
if (match == null) {
return false;
}
for (char aMatch : match) {
if (c == aMatch) {
return true;
}
}
return false;
}
public static int findFirstEqual(char[] source, int index, char[] match) {
if (source == null || match == null) {
return -1;
}
for (int i = index, length = source.length; i < length; i++) {
if (equalsOne(source[i], match)) {
return i;
}
}
return -1;
}
public static int findFirstEqual(char[] source, int index, char match) {
if (source == null) {
return -1;
}
for (int i = index; i < source.length; i++) {
if (source[i] == match) {
return i;
}
}
return -1;
}
public static int findFirstDiff(char[] source, int index, char[] match) {
if (source == null || match == null) {
return -1;
}
for (int i = index; i < source.length; i++) {
if (!equalsOne(source[i], match)) {
return i;
}
}
return -1;
}
public static int findFirstDiff(char[] source, int index, char match) {
if (source == null) {
return -1;
}
for (int i = index; i < source.length; i++) {
if (source[i] != match) {
return i;
}
}
return -1;
}
public static boolean isLowercaseLetter(char c) {
return (c >= 'a') && (c <= 'z');
}
public static boolean isUppercaseLetter(char c) {
return (c >= 'A') && (c <= 'Z');
}
public static boolean isLetter(char c) {
return ((c >= 'a') && (c <= 'z')) || ((c >= 'A') && (c <= 'Z'));
}
public static boolean isDigit(char c) {
return (c >= '0') && (c <= '9');
}
public static boolean isLetterOrDigit(char c) {
return isDigit(c) || isLetter(c);
}
public static boolean isWordChar(char c) {
return isDigit(c) || isLetter(c) || (c == '_');
}
public static boolean isPropertyNameChar(char c) {
return isDigit(c) || isLetter(c) || (c == '_') || (c == '.') || (c == '[') || (c == ']');
}
public static boolean isAscii(char ch) {
return ch < 128;
}
public static boolean isAsciiPrintable(char ch) {
return ch >= 32 && ch < 127;
}
public static boolean isAsciiControl(char ch) {
return ch < 32 || ch == 127;
}
public static boolean isChinese(char c) {
return (c >= 0x4e00 && c <= 0x9fa5);
}
public static char toUpperAscii(char c) {
if (isLowercaseLetter(c)) {
c -= (char) 0x20;
}
return c;
}
public static char toLowerAscii(char c) {
if (isUppercaseLetter(c)) {
c += (char) 0x20;
}
return c;
}
}
© 2015 - 2025 Weber Informatics LLC | Privacy Policy