objective-C中的序列化(serialize)与反序列化(deserialize)

obj-c中有一类对象：NSArray,NSDictionary,NSString,NSNumber,NSDate,NSData以及它们的可变版本(指NSMutableArray,NSMutableDictionary...这一类) ，都可以方便的将自身的数据以某种格式(比如xml格式)序列化后保存成本地文件。

示例代码：NSArrayTest.h

#import <Foundation/Foundation.h>
#define FILE_NAME @"/tmp/data.txt"

@interface NSArrayTest : NSObject {	
}

-(void) Test;

@end

NSArrayTest.m

#import "NSArrayTest.h"

@implementation NSArrayTest

-(void) Test
{
	NSArray *arr = [NSArray arrayWithObjects:@"one",@"two",@"three",nil];//注:最后一个要以nil结尾
	[arr writeToFile:FILE_NAME atomically:YES];//(序列化为xml格式后)保存文件
	
	NSArray *arr2 = [NSArray arrayWithContentsOfFile:FILE_NAME];//read file
	NSLog(@"%@",arr2);

}

@end

运行结果：

2011-03-03 14:20:01.501 pList[1246:a0f] (
    one,
    two,
    three
)

如果查看/tmp/data.txt，能看到下面的内容：

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<array>
	<string>one</string>
	<string>two</string>
	<string>three</string>
</array>
</plist>

即NSArray默认是以xml格式来序列化对象的.

如果你用来存放数据的类是自己定义的，并不是上面这些预置的对象，那么就要借助正式协议NSCoding来实现序列化和反序列化。

比如，我们有一个自己的类Sample.h

#import <Foundation/Foundation.h>

@interface Sample : NSObject<NSCoding> {

	NSString* name;
	int magicNumber;
	float shoeSize;
	NSMutableArray *subThingies;	
}

@property(copy) NSString* name;
@property int magicNumber;
@property float shoeSize;
@property (retain) NSMutableArray *subThingies;


-(id) initWithName:(NSString *)n magicNumber:(int)m shoeSize:(float) ss;

@end

这里我们定义几个不同类型的属性，有字符串，有整数，有浮点数，还有一个可变长的数组对象

Sample.m

#import "Sample.h"

@implementation Sample

@synthesize name;
@synthesize magicNumber;
@synthesize shoeSize;
@synthesize subThingies;

-(id) initWithName:(NSString *)n magicNumber:(int)m shoeSize:(float)ss
{
	if (self=[super init])
	{
		self.name = n;
		self.magicNumber = m;
		self.shoeSize = ss;
		self.subThingies = [NSMutableArray array];
	}
	return (self);
}

-(void) dealloc
{
	[name release];
	[subThingies release];
	[super dealloc];
}

//将对象编码(即:序列化)
-(void) encodeWithCoder:(NSCoder *)aCoder
{
	[aCoder encodeObject:name forKey:@"name"];
	[aCoder encodeInt:magicNumber forKey:@"magicNumber"];
	[aCoder	encodeFloat:shoeSize forKey:@"shoeSize"];
	[aCoder	encodeObject:subThingies forKey:@"subThingies"];
}

//将对象解码(反序列化)
-(id) initWithCoder:(NSCoder *)aDecoder
{
	if (self=[super init])
	{
		self.name = [aDecoder decodeObjectForKey:@"name"];
		self.magicNumber = [aDecoder decodeIntForKey:@"magicNumber"];
		self.shoeSize = [aDecoder decodeFloatForKey:@"shoeSize"];
		self.subThingies = [aDecoder decodeObjectForKey:@"subThingies"];
	}
	return (self);
	
}


-(NSString*) description
{
	NSString *description = [NSString stringWithFormat:@"%@:%d/%.1f %@",name,magicNumber,shoeSize,subThingies];
	return (description);
}


@end

注意其中的：encodeWithCoder与initWithCoder，这是NSCoding协议中定义的二个方法，用来实现对象的编码与解码。其实现也不复杂，利用的是key-value的经典哈希结构。当然一般在编码中，对于key的名字字符串，建议用define以常量方式事先定义好，以避免开发人员字符串键入错误。

测试一下：

#import <Foundation/Foundation.h>
#import "Sample.h"

int main (int argc, const char * argv[]) {
    NSAutoreleasePool * pool = [[NSAutoreleasePool alloc] init];	

	Sample *s1 = [[Sample alloc] initWithName:@"thing1" magicNumber:42 shoeSize:10.5];	
	[s1.subThingies addObject:@"1"];
	[s1.subThingies addObject:@"2"];
	NSLog(@"%@",s1);	
	NSData	*data1 = [NSKeyedArchiver archivedDataWithRootObject:s1];//将s1序列化后,保存到NSData中
	[s1 release];
	[data1 writeToFile:@"/tmp/data.txt" atomically:YES];//持久化保存成物理文件
			
	NSData *data2 = [NSData dataWithContentsOfFile:@"/tmp/data.txt"];//读取文件	
	Sample *s2 = [NSKeyedUnarchiver unarchiveObjectWithData:data2];//反序列化	
    NSLog(@"%@",s2);
	 
	[pool drain];
    return 0;
}

运行结果：

2011-03-03 14:36:48.540 pList[1322:a0f] thing1:42/10.5 (
    1,
    2
)
2011-03-03 14:36:48.548 pList[1322:a0f] thing1:42/10.5 (
    1,
    2
)

查看/tmp/data.txt，能看到以下内容：

由于经过了编码，里面的内容没有象前面的NSArray那样可读性强。