What are advantages and disadvantages of union in c language?

Answer 1

A union in C++ is similar to a C-style union in that the members of the union are all mapped to the same memory address. In other words, if you were to unify a long with a char, the char would be mapped to the first byte of the long. However, unlike C unions, C++ unions can be derived from other unions, structs or classes, as well as act as base classes for derived unions, structs and classes.

Answer 2

The basic advantage is that union will use the memory space of the datatype which has the highest memory....

hence memory consumption will be less...

But when u use structure the total memory will be the sum of the memory of all datatypes.. ie.(higher memory allocation)

Disadvantage is that..... when you use the union.... only the last entered variable can be directly accessed as only the last added data to the union will exist in the memory...

Answer 3

We use a union whenever we need to select between two or more representations where only one representation is in use (active) at any given time. All members of a union share the same memory address, thus a union provides a convenient means of reducing memory consumption when we require alternate representations. The length of a union is equal to the longest member (in bytes) but, otherwise, all members behave according to their respective types.

The main disadvantage is that there's nothing to prevent a programmer from writing to one member and subsequently reading another member. Although this unavoidable shortcoming is often used to implement a cast between member types that share a common representation (a 32-bit integer and a 4-byte signed char array, for instance), it's tempting to use this technique to forcefully cast between types that do not share a common representation (an int and a pointer to int would be an unsafe cast, for instance). In short, use unions for the purpose intended. If you have to shortcut the built-in casts using union members, you should seriously reconsider your design choices.

For members that do not share a common representation, unions must be used in conjunction with a type field to determine which member is currently active. In some cases, the type field can be dual purpose and therefore does not consume any additional memory (after all, the intent is to save memory), however where there is a cost, it needn't be expensive. Most unions have relatively few members such that an enumerated char type generally suffices for a type field. Typically, the union and its type field will be members of the same struct.