Java代码如何反序列化PHP序列化数组后的字符串

刚进公司第一个任务是写一个导出财务报表的功能：按照财务那边的需求，从数据库查询指定的字段导出到Excel表格，方便财务对账。公司以前的系统是用PHP写的，有些数据库字段是通过PHP的serialize()方法序列化之后存进数据库的，如下所示

a:2:{i:0;a:2:{s:7:"orderid";s:2:"62";s:5:"level";i:1;}i:1;a:2:{s:7:"orderid";s:2:"74";s:5:"level";i:1;}}

PHP可以通过unserialize()反序列化解析出需要的内容，上面的字符串反序列化之后其实是一个二维数组。那么问题来了，我不会写PHP，只能用Java的方法去解析这段字符串。

一看到这个字符串，我以为是JSON格式的字符串，或许可以通过JSON方法解析；后来认真看，这不是JSON格式，后来又想，能不能写个通用方法，把这种类型的字符串转换成JSON格式，再通过JSON方法解析，试了一下，挺麻烦的，不好实现。所以最后想，网上有没有现成Java的方法反序列化的工具类，然后解析出需要的内容。

如何用Java实现反序列化呢？我在网上查了很多资料，找了很久终于找到一个比较靠谱的方法。在项目中导入以下两个类：

PHPSerialize类：

import java.util.regex.*;
import java.util.*;

public class PHPSerialize {
    /**
     * To store partially decomposed serialized string
     */
    private StringBuffer sb;

    /** Creates a new instance of phpserialize */
    public PHPSerialize() {
    }

    /**
     * Unserializes a single value into a PHPValue
     * 
     * @param String
     *            s String to unserialize
     * @return PHPValue
     */
    public PHPValue unserialize(String s) {
        sb = new StringBuffer(s);
        PHPValue phpvalue = getNext();
        // println("value = " + phpvalue.toString());
        return phpvalue;
    }

    /**
     * Get the next token in the serialized string
     * 
     * 
     */
    private PHPValue getNext() {
        // println("getNext = "+ sb.toString());
        PHPValue phpvalue;
        Object value = new Object();
        // Current character
        String c;
        // What type of value is this
        String what = "";
        // Consume & throw away characters
        int consumeChars = 0;
        // Length of the string/array
        int len = -1;
        // Buffer to read in string
        StringBuffer tmp = new StringBuffer();
        // length of sb is constantly changing.
        while (sb.length() > 0) {
            c = sb.substring(0, 1);
            sb = sb.deleteCharAt(0);
            // println(c +"\twhat="+what+"\ttmp="+tmp.toString());
            // Ignore chars
            if (consumeChars-- > 0) {
                continue;
            }
            /**
             * This is a hackish pseudo state machine 'what' is the state
             * 
             */
            if (what.equals("array_length")) {
                if (c.equals(":")) {
                    len = Integer.parseInt(tmp.toString());
                    tmp.delete(0, tmp.length());
                    // println("\tarray_length="+ len);
                    what = "array_value";
                }
                if (len == -1) {
                    tmp.append(c);
                }
            } else if (what.equals("array_value")) {
                // value = new PHPValue( new HashMap() );
                HashMap hash = new HashMap();
                for (int i = 0; i < len; i++) {
                    PHPValue hashKey = getNext();
                    PHPValue hashValue = getNext();
                    // Force the keys of the HashMap to be type String
                    hash.put(hashKey.value, hashValue);
                }
                value = hash;
                // return new PHPValue(value);
                what = "set";
            } else if (what.equals("string_length")) {
                if (c.equals(":")) {
                    len = Integer.parseInt(tmp.toString());
                    tmp.delete(0, tmp.length());
                    // println("\tstring_length="+ len);
                    what = "string_value";
                    consumeChars = 1;
                }
                if (len == -1) {
                    tmp.append(c);
                }
            } else if (what.equals("string_value")) {
                if (len-- > 0) {
                    tmp.append(c);
                } else {
                    value = tmp.toString();
                    what = "set";
                    // println("\tstring_value="+ value);
                }
            } else if (what.equals("integer")) {
                if (c.equals(";")) {
                    value = tmp.toString();
                    what = "set";
                } else {
                    tmp.append(c);
                }
            } else if (what.equals("double")) {
                if (c.equals(";")) {
                    value = tmp.toString();
                    what = "set";
                } else {
                    tmp.append(c);
                }
            } else if (what.equals("null")) {
                if (c.equals(";")) {
                    value = null;
                    what = "set";
                } else {
                    // Bad
                }
            }
            // What kind of variable is coming up next
            else {
                if (c.equals("a")) {
                    what = "array_length";
                } else if (c.equals("s")) {
                    what = "string_length";
                } else if (c.equals("i")) {
                    what = "integer";
                } else if (c.equals("d")) {
                    what = "double";
                } else if (c.equals("n")) {
                    what = "null";
                    // do NOT consume the next char ";"
                    continue;
                } else {
                    continue;
                }

                // Consume the ":" after variable type declaration
                consumeChars = 1;
                continue;
            }

            if (what.equals("set")) {
                return new PHPValue(value);
            }

        }

        return new PHPValue(value);
    }

}

PHPValue类：

import java.util.HashMap;

/**
 * Since i have to declare the return type of a function, and i have no idea
 * ahead of time what value is being unserialized, this helper class is a hack
 * so that i can return any variable type as PHPValue and deal with it later.
 * 
 * @author Scott Hurring [scott at hurring dot com]
 * @version v0.01a ALPHA!!
 * @licence GNU GPL
 */
public class PHPValue {
    /**
     * The raw object given to us by unserialize()
     */
    private Object obj = null;
    private HashMap hobj = new HashMap();

    /**
     * The toString() value of the Object passed into PHPValue
     */
    public String value = "";

    /**
     * Is the PHPValue one of the following?
     */
    public boolean isNull = false;
    public boolean isInteger = false;
    public boolean isDouble = false;
    public boolean isString = false;
    public boolean isHashMap = false;

    /**
     * Pass in an object, and PHPValue: 1. Attempts to try and figure out what
     * it is 2. Stringifies it 3. Sets a boolean flag to let everyone know what
     * we think it is.
     * 
     */
    public PHPValue(Object o) {
        obj = o;

        if (obj == null) {
            isNull = true;
            value = "";
        } else if (obj.getClass().toString().equals("class java.lang.Integer")) {
            isInteger = true;
            value = ((Integer) obj).toString();
        } else if (obj.getClass().toString().equals("class java.lang.Double")) {
            isDouble = true;
            value = ((Double) obj).toString();
        } else if (obj.getClass().toString().equals("class java.util.HashMap")) {
            isHashMap = true;
            value = obj.toString();
        } else if (obj.getClass().toString().equals("class java.lang.String")) {
            isString = true;
            value = obj.toString();
        } else {
            value = obj.toString();
        }

        // debug
        if (obj != null) {
            // System.out.println("PHPValue: '" + obj.getClass() + "' = " +
            // value);
        }

    }

    /**
     * Convert PHPValue into a double
     * 
     * @return double
     */
    public double toDouble() {
        if (isDouble) {
            return Double.parseDouble(value);
        } else {
            return 0.0;
        }
    }

    /**
     * Convert PHPValue into an integer
     * 
     * @return int
     */
    public int toInteger() {
        if (isInteger) {
            return Integer.parseInt(value);
        } else {
            return 0;
        }
    }

    /**
     * PHPValue's string-ified value
     * 
     * @return String
     */
    public String toString() {
        if (!isNull) {
            return value;
        } else {
            return "";
        }
    }

    /**
     * Cast PHPValue into HashMap
     * 
     * @return HashMap
     */
    public HashMap toHashMap() {
        return (HashMap) obj;
    }

    /**
     * Return the raw object
     * 
     * @return double
     */
    public Object toObject() {
        return obj;
    }

}

如何调用：

public static void main(String[] args) {
        String str = "a:2:{i:0;a:2:{s:7:\"orderid\";s:2:\"77\";s:5:\"level\";i:1;}i:1;"
                    + "a:2:{s:7:\"orderid\";s:2:\"78\";s:5:\"level\";i:1;}}";
        PHPSerialize p = new PHPSerialize();
        PHPValue c = p.unserialize(str);
        System.out.println(c.toHashMap().toString());
}

反序列化出来的结果如下所示，然后你再利用字符串拆分出所需内容。

{0={orderid=77, level=1}, 1={orderid=78, level=1}}

不过上面的方法存在一个问题，就是要解析的内容存在中文时就会报错，如下：

a:2:{s:16:"carrier_realname";s:7:"测试2";s:14:"carrier_mobile";s:11:"18725195368";}

我的解决办法是直接拆分这个字符串，一步一步解析出所需的内容。因为这个字符串的格式的固定的，有一点的规律，所以就可以写一个方法：

//得到有关收货人的字符串后，根据特定规则解析字符串，得到所需内容
                    if (o_carrier.length() > 22) {
                        //收货人
                        String carrier_realname = o_carrier.substring(
                                o_carrier.indexOf(":\"carrier_realname\";s:") + 22,
                                o_carrier.indexOf(":\"carrier_mobile\";s:"));
                        //联系电话
                        String carrier_mobile = o_carrier.substring(o_carrier.indexOf(":\"carrier_mobile\";s:") + 20,
                                o_carrier.indexOf("\";}"));

                        //取出两个"号中间的字符内容
                        String regex = "\"(.*)\"";
                        Pattern pattern = Pattern.compile(regex);
                        Matcher matcher = pattern.matcher(carrier_realname);
                        while (matcher.find()) {
                            //收货人
                            System.out.println(matcher.group(1));
                        }

                        //根据"号拆分字符串，存放到字符串数组
                        String mobile[] = carrier_mobile.split("\"");
                        //联系电话
                        System.out.println(mobile[1]);
                    }