14.11. Object Serialization

This chapter and Chapter 13 introduced the object-oriented style of input/output. However, our examples concentrated on I/O of fundamental types rather than objects of user-defined types. In Chapter 10, we showed how to input and output objects using operator overloading. We accomplished object input by overloading the stream extraction operator, >>, for the appropriate istream. We accomplished object output by overloading the stream insertion operator, <<, for the appropriate ostream. In both cases, only an object’s data members were input or output, and, in each case, they were in a format meaningful only for objects of that particular type. An object’s member functions are not input or output with the object’s data; rather, one copy of the class’s member functions remains available internally and is shared by all objects of the class.

When object data members are output to a disk file, we lose the object’s type information. We store only the values of the object’s attributes, not type information, on the disk. If the program that reads this data knows the object type to which the data corresponds, the program can read the data into an object of that type as we did in our random-access file examples.

An interesting problem occurs when we store objects of different types in the same file. How can we distinguish them (or their collections of data members) as we read them into a program? The problem is that objects typically do not have type fields (we discussed this issue in Chapter 12).

One approach used by several programming languages is called object serialization. A so-called serialized object is an object represented as a sequence of bytes that includes the object’s data as well as information about the object’s type and the types of data stored in the object. After a serialized object has been written to a file, it can be read from the file and deserialized—that is, the type information and bytes that represent the object and its data can be used to recreate the object in memory. C++ does not provide a built-in serialization mechanism; however, there are third party and open source C++ libraries that support object serialization. The open source Boost C++ Libraries (www.boost.org) provide support for serializing objects in text, binary and extensible markup language (XML) formats (www.boost.org/libs/serialization/doc/index.html).