KickJava   Java API By Example, From Geeks To Geeks.

Java > Open Source Codes > com > quadcap > util > text > Soundex


1 package com.quadcap.util.text;
2
3 /* Copyright 1997 - 2003 Quadcap Software. All rights reserved.
4  *
5  * This software is distributed under the Quadcap Free Software License.
6  * This software may be used or modified for any purpose, personal or
7  * commercial. Open Source redistributions are permitted. Commercial
8  * redistribution of larger works derived from, or works which bundle
9  * this software requires a "Commercial Redistribution License"; see
10  * http://www.quadcap.com/purchase.
11  *
12  * Redistributions qualify as "Open Source" under one of the following terms:
13  *
14  * Redistributions are made at no charge beyond the reasonable cost of
15  * materials and delivery.
16  *
17  * Redistributions are accompanied by a copy of the Source Code or by an
18  * irrevocable offer to provide a copy of the Source Code for up to three
19  * years at the cost of materials and delivery. Such redistributions
20  * must allow further use, modification, and redistribution of the Source
21  * Code under substantially the same terms as this license.
22  *
23  * Redistributions of source code must retain the copyright notices as they
24  * appear in each source code file, these license terms, and the
25  * disclaimer/limitation of liability set forth as paragraph 6 below.
26  *
27  * Redistributions in binary form must reproduce this Copyright Notice,
28  * these license terms, and the disclaimer/limitation of liability set
29  * forth as paragraph 6 below, in the documentation and/or other materials
30  * provided with the distribution.
31  *
32  * The Software is provided on an "AS IS" basis. No warranty is
33  * provided that the Software is free of defects, or fit for a
34  * particular purpose.
35  *
36  * Limitation of Liability. Quadcap Software shall not be liable
37  * for any damages suffered by the Licensee or any third party resulting
38  * from use of the Software.
39  */
40
41 /**
42  * SOUNDEX Utilities.
43  *
44  * @author Stan Bailes
45  */
46 public class Soundex {
47     static final OctetMap alpha = new OctetMap('a', 'z');
48     static final OctetMap Alpha = new OctetMap('A', 'Z');
49     static {
50     Alpha.include('a', 'z');
51     }
52
53     /* ABCDEFGHIJKLMNOPQRSTUVWXYZ */
54     static final String JavaDoc sMap = "01230120022455012623010202";
55     static final char scode(int c) {
56     if (alpha.has(c)) {
57         return sMap.charAt(c - 'a');
58     } else {
59         return sMap.charAt(c - 'A');
60     }
61     }
62
63     public static final String JavaDoc soundex(String JavaDoc s) {
64     char[] ret = new char[4];
65     char last = 'x';
66     int pos = 0;
67     for (int i = 0; i < s.length() && pos < 4; i++) {
68         int c = s.charAt(i) & 0xff;
69         if (Alpha.has(c)) {
70         if (pos == 0) {
71             ret[pos++] = Character.toUpperCase((char)c);
72         } else {
73             char code = scode(c);
74             if (code != '0' && code != last) {
75             ret[pos++] = code;
76             last = code;
77             }
78         }
79         }
80     }
81     if (pos == 0) return "";
82     while (pos < 4) ret[pos++] = '0';
83     return new String JavaDoc(ret);
84     }
85
86     public static final int difference(String JavaDoc a, String JavaDoc b) {
87     String JavaDoc sa = soundex(a);
88     String JavaDoc sb = soundex(b);
89     int diff = 0;
90     for (int i = 0; i < 4; i++) {
91         if (sa.charAt(i) == sb.charAt(i)) diff++;
92     }
93     return diff;
94     }
95
96     //#ifndef RELEASE
97 static String JavaDoc[] data = {
98     "blather", "blabber", "Smith", "Smyth",
99     "abcdefghijklmnopqrstuvwxyz", "a", "b", "bed", "BBD",
100     "ABCDEFGHIJKLMNOPQRSTUVWXYZ", "smithers", "smothers", "brothers"
101     };
102     public static void main(String JavaDoc[] args) {
103     for (int i = 0; i < data.length; i++) {
104         System.out.println(soundex(data[i]) + ": " + data[i]);
105     }
106     }
107     //#endif
108 }
109
Popular Tags