ChatGPT解决这个技术问题 Extra ChatGPT

What are the differences between struct and class in C++?

This question was already asked in the context of C#/.Net.

Now I'd like to learn the differences between a struct and a class in C++. Please discuss the technical differences as well as reasons for choosing one or the other in OO design.

I'll start with an obvious difference:

If you don't specify public: or private:, members of a struct are public by default; members of a class are private by default.

I'm sure there are other differences to be found in the obscure corners of the C++ specification.

This link summarizes the differences well.
Why people all use struct to build a tree then? Because it seems that the difference is not that much. BTW, That is a great website.@sjsam
Looking for the difference between struct in C and C++? See here.
@JW.ZG Not "all" people do! Those who do that either simply prefer to, or do not realise what struct means in C++. ;) But there's no reason you can't use the class keyword instead.

C
Cheers and hth. - Alf

You forget the tricky 2nd difference between classes and structs.

Quoth the standard (§11.2.2 in C++98 through C++11):

In absence of an access-specifier for a base class, public is assumed when the derived class is declared struct and private is assumed when the class is declared class.

And just for completeness' sake, the more widely known difference between class and struct is defined in (11.2):

Member of a class defined with the keyword class are private by default. Members of a class defined with the keywords struct or union are public by default.

Additional difference: the keyword class can be used to declare template parameters, while the struct keyword cannot be so used.


That's not really a second difference, it's just that the question stated the difference incompletely. All subobjects are private by default when using the class, public by default with struct, and both define a class type.
I think you missed the point, Ben. It's the inheritance that's public for structs and private for classes. That's a very important distinction, and completely unrelated to actual member access. For example, the language could very well have defined it otherwise, e.g. both are publicly inherited by default, and you would get very different implications.
Actually, the real tricky difference between struct and class is that the latter can be used in place of a typename to declare a template parameter, while the former cannot. :)
@MrUniverse: This is utterly wrong. (There's a school that uses this for a convention, but it's not syntactically enforced.)
@sbi "the real tricky difference between struct and class" -- Cute, but the question is about the difference between a struct and a class, not about differences in where the keywords can be used.
d
dgellow

Quoting The C++ FAQ,

[7.8] What's the difference between the keywords struct and class? The members and base classes of a struct are public by default, while in class, they default to private. Note: you should make your base classes explicitly public, private, or protected, rather than relying on the defaults. Struct and class are otherwise functionally equivalent. OK, enough of that squeaky clean techno talk. Emotionally, most developers make a strong distinction between a class and a struct. A struct simply feels like an open pile of bits with very little in the way of encapsulation or functionality. A class feels like a living and responsible member of society with intelligent services, a strong encapsulation barrier, and a well defined interface. Since that's the connotation most people already have, you should probably use the struct keyword if you have a class that has very few methods and has public data (such things do exist in well designed systems!), but otherwise you should probably use the class keyword.


J
Jon Hanna

It's worth remembering C++'s origins in, and compatibility with, C.

C has structs, it has no concept of encapsulation, so everything is public.

Being public by default is generally considered a bad idea when taking an object-oriented approach, so in making a form of C that is natively conducive to OOP (you can do OO in C, but it won't help you) which was the idea in C++ (originally "C With Classes"), it makes sense to make members private by default.

On the other hand, if Stroustrup had changed the semantics of struct so that its members were private by default, it would have broken compatibility (it is no longer as often true as the standards diverged, but all valid C programs were also valid C++ programs, which had a big effect on giving C++ a foothold).

So a new keyword, class was introduced to be exactly like a struct, but private by default.

If C++ had come from scratch, with no history, then it would probably have only one such keyword. It also probably wouldn't have made the impact it made.

In general, people will tend to use struct when they are doing something like how structs are used in C; public members, no constructor (as long as it isn't in a union, you can have constructors in structs, just like with classes, but people tend not to), no virtual methods, etc. Since languages are as much to communicate with people reading the code as to instruct machines (or else we'd stick with assembly and raw VM opcodes) it's a good idea to stick with that.


It's probably also worth noting that it was probably easier to have classes and structures work the same under the hood, than to have them be fundamentally different. There might in retrospect have been some benefit to e.g. specifying that anything declared as a struct must be a PODS (i.e. forbidding structs from having virtual members, non-trivial default constructors or destructors, etc.) but such a thing has not to my knowledge ever been required by any C++ standard nor implementation.
@supercat the fact that it was an OO language in which a vast repertoire of programs written in a non-OO language were valid programs, brought the concept to the fore, and it does seem to come up more there than in more purely OO languages (and of course is actually defined in the standard). The history of early C++ definitely explains a lot about it. Stroustrup's "The Design and Evolution of C++" is an interesting read.
I perceive a religious war between people who feel that everything which isn't "oop-ish", should be, and those who think that OOP should be regarded as a tool, but not the tool. C++ certainly seems to embrace the latter philosophy, but I really don't know much about languages before it. My own philosophy is that when one wants an object one should use an object, and when one wants an aggregation of related but independent variables, one should use that. C++ makes no syntactic distinction between those kinds of things, and .NET tries to blur that distinction, but the little I know of D...
...suggests that it recognizes it moreso than C++ or .NET. Aggregations of variables aren't exactly OOPish, but that doesn't mean they shouldn't be employed when they're useful.
I disagree with "C has structs, it has no concept of encapsulation, so everything is public.", You can hide a definition of a struct inside a *.c-File and let other translations units only use it with a pointer. So you have a even stronger encapsulation, because other translations units do not even know what is inside the struct.
C
Community

Class' members are private by default. Struct's members are public by default. Besides that there are no other differences. Also see this question.


Not only members, but all access, including inheritance.
Well, I'd say the other difference is semantics. Struct to many people makes them think "data structure" as a leftover from C.
@KrisKumler: Those are not "semantics"; they are personal feelings. Both struct and class declare a class with identical semantics (notwithstanding parse-wise interpretation of the definition body).
c
crashmstr

According to Stroustrup in the C++ Programming Language:

Which style you use depends on circumstances and taste. I usually prefer to use struct for classes that have all data public. I think of such classes as "not quite proper types, just data structures."

Functionally, there is no difference other than the public / private


What difference? Please explain.
d
dgellow

Members of a class are private by default and members of struct are public by default.

For example program 1 fails in compilation and program 2 works fine.

// Program 1
#include <stdio.h>
 
class Test {
    int x; // x is private
};

int main()
{
  Test t;
  t.x = 20; // compiler error because x is private
  getchar();
  return 0;
}
// Program 2
#include <stdio.h>
 
struct Test {
    int x; // x is public
};

int main()
{
  Test t;
  t.x = 20; // works fine because x is public
  getchar();
  return 0;
}

When deriving a struct from a class/struct, default access-specifier for a base class/struct is public. And when deriving a class, default access specifier is private.

For example program 3 fails in compilation and program 4 works fine.

// Program 3
#include <stdio.h>
 
class Base {
public:
    int x;
};
 
class Derived : Base { }; // is equivalent to class Derived : private Base {}
 
int main()
{
  Derived d;
  d.x = 20; // compiler error because inheritance is private
  getchar();
  return 0;
}
// Program 4
#include <stdio.h>
 
class Base {
public:
    int x;
};
 
struct Derived : Base { }; // is equivalent to struct Derived : public Base {}
 
int main()
{
  Derived d;
  d.x = 20; // works fine because inheritance is public
  getchar();
  return 0;
}

I have removed code examples 3-7 as mentioned by experts
6
64BitBob

STRUCT is a type of Abstract Data Type that divides up a given chunk of memory according to the structure specification. Structs are particularly useful in file serialization/deserialization as the structure can often be written to the file verbatim. (i.e. Obtain a pointer to the struct, use the SIZE macro to compute the number of bytes to copy, then move the data in or out of the struct.)

Classes are a different type of abstract data type that attempt to ensure information hiding. Internally, there can be a variety of machinations, methods, temp variables, state variables. etc. that are all used to present a consistent API to any code which wishes to use the class.

In effect, structs are about data, classes are about code.

However, you do need to understand that these are merely abstractions. It's perfectly possible to create structs that look a lot like classes and classes that look a lot like structs. In fact, the earliest C++ compilers were merely pre-compilers that translates C++ code to C. Thus these abstractions are a benefit to logical thinking, not necessarily an asset to the computer itself.

Beyond the fact that each is a different type of abstraction, Classes provide solutions to the C code naming puzzle. Since you can't have more than one function exposed with the same name, developers used to follow a pattern of _(). e.g. mathlibextreme_max(). By grouping APIs into classes, similar functions (here we call them "methods") can be grouped together and protected from the naming of methods in other classes. This allows the programmer to organize his code better and increase code reuse. In theory, at least.


This is the most non-obvious thing about structs and classes. "Structs are about data, classes are about code" can be rephrased as "structs are about data, classes are about data, security and operations performed on this data"
It is not possible to create a struct in C++. It is only possible to create a class using the inherited struct keyword.
This answer doesn't seem to be about C++. In C++, struct declares classes (including public/private/protected, inheritance, etc).
How are structs abstract? Writing raw structs to a file is fraught with problems even in C (the usual problems include padding, endianness, different word size, etc). What is this SIZE macro you're talking about?
T
Tim Cooper

The only other difference is the default inheritance of classes and structs, which, unsurprisingly, is private and public respectively.


4
463035818_is_not_a_number

The difference between class and struct is a difference between keywords, not between data types. This two

struct foo : foo_base { int x;};
class bar : bar_base { int x; };

both define a class type. The difference of the keywords in this context is the different default access:

foo::x is public and foo_base is inherited publicly

bar::x is private and bar_base is inherited privately


m
melpomene

The members of a structure are public by default, the members of class are private by default. Default inheritance for Structure from another structure or class is public.Default inheritance for class from another structure or class is private.

class A{    
public:    
    int i;      
};

class A2:A{    
};

struct A3:A{    
};


struct abc{    
    int i;
};

struct abc2:abc{    
};

class abc3:abc{
};


int _tmain(int argc, _TCHAR* argv[])
{    
    abc2 objabc;
    objabc.i = 10;

    A3 ob;
    ob.i = 10;

    //A2 obja; //privately inherited
    //obja.i = 10;

    //abc3 obss;
    //obss.i = 10;
}

This is on VS2005.


_tmain is not standard C++.
M
MSalters

Not in the specification, no. The main difference is in programmer expectations when they read your code in 2 years. structs are often assumed to be POD. Structs are also used in template metaprogramming when you're defining a type for purposes other than defining objects.


K
KPexEA

One other thing to note, if you updated a legacy app that had structs to use classes you might run into the following issue:

Old code has structs, code was cleaned up and these changed to classes. A virtual function or two was then added to the new updated class.

When virtual functions are in classes then internally the compiler will add extra pointer to the class data to point to the functions.

How this would break old legacy code is if in the old code somewhere the struct was cleared using memfill to clear it all to zeros, this would stomp the extra pointer data as well.


Code that uses memfill to clear a struct probably has other offensive habits. Go with caution.
@DavidThornley: Use of a zero-fill to erase or initialize a data structure is standard in C. In a mixed-language project, one cannot expect the code in the C part to avoid calloc in favor of new. All one can do is try to ensure that any structures used by the C part of the code are in fact PODS.
S
Stefan Popescu

Another main difference is when it comes to Templates. As far as I know, you may use a class when you define a template but NOT a struct.

template<class T> // OK
template<struct T> // ERROR, struct not allowed here

If you have intended to emphasise that structures cannot be used as generalized types, then you a wrong. And struct is not allowed there just by convention (better say it's class keyword that is used in different sense here, see Use 'class' or 'typename' for template parameters).
The keyword isn't allowed there, but you can pass in a class you defined with struct just fine. Simply try it. struct T{}; template <typename T> void foo() {} foo<T>();
B
Bathsheba

Member of a class defined with the keyword class are private by default. Members of a class defined with the keywords struct (or union) are public by default. In absence of an access-specifier for a base class, public is assumed when the derived class is declared struct and private is assumed when the class is declared class. You can declare an enum class but not an enum struct. You can use template but not template.

Note also that the C++ standard allows you to forward-declare a type as a struct, and then use class when declaring the type and vice-versa. Also, std::is_class<Y>::value is true for Y being a struct and a class, but is false for an enum class.


nice answer, though i am missing a clear unmistakable 0) the difference between struct and class in c++ is a difference between keywords, not between data types (or some better alternative ;)
@formerlyknownas_463035818: But don't you think std::is_class<Y>::value covers that?
yeah, just thought it could be more prominent, because thats the common misunderstanding and once thats out of the way the rest is just corollaries.
Nevermind, your answer is perfect as is, i wrote another one, doesnt hurt to have one more
n
nutario

Here is a good explanation: http://carcino.gen.nz/tech/cpp/struct_vs_class.php

So, one more time: in C++, a struct is identical to a class except that the members of a struct have public visibility by default, but the members of a class have private visibility by default.


f
finnw

It's just a convention. Structs can be created to hold simple data but later evolve time with the addition of member functions and constructors. On the other hand it's unusual to see anything other than public: access in a struct.


G
Gregory Pakosz

ISO IEC 14882-2003 9 Classes §3 A structure is a class defined with the class-key struct; its members and base classes (clause 10) are public by default (clause 11).


y
ymett

The other answers have mentioned the private/public defaults, (but note that a struct is a class is a struct; they are not two different items, just two ways of defining the same item).

What might be interesting to note (particularly since the asker is likely to be using MSVC++ since he mentions "unmanaged" C++) is that Visual C++ complains under certain circumstances if a class is declared with class and then defined with struct (or possibly the other way round), although the standard says that is perfectly legal.


I know it's a very late comment, but be aware that the aforementioned warning is not without merit as the Windows ABI mangles structs and classes differently and therefore mixing (forward) declarations of class and struct can actually lead to hard to comprehend linking problems!
y
yshivakathik

. In classes all the members by default are private but in structure members are public by default. There is no term like constructor and destructor for structs, but for class compiler creates default if you don't provide. Sizeof empty structure is 0 Bytes wer as Sizeof empty class is 1 Byte The struct default access type is public. A struct should typically be used for grouping data. The class default access type is private, and the default mode for inheritance is private. A class should be used for grouping data and methods that operate on that data. In short, the convention is to use struct when the purpose is to group data, and use classes when we require data abstraction and, perhaps inheritance. In C++ structures and classes are passed by value, unless explicitly de-referenced. In other languages classes and structures may have distinct semantics - ie. objects (instances of classes) may be passed by reference and structures may be passed by value. Note: There are comments associated with this question. See the discussion page to add to the conversation.

There is no term like constructor and destructor for structs, but for class compiler creates default if you don't provide.

Sizeof empty structure is 0 Bytes wer as Sizeof empty class is 1 Byte The struct default access type is public. A struct should typically be used for grouping data.


"2. Sizeof empty structure is 0 Bytes wer as Sizeof empty class is 1 Byte". False. I just tested with g++. You might be thinking of Empty Base Optimization, but that applies to both too.
Not true. In C++ the size of an empty struct is 1. In C an empty struct is not allowed. GCC allows empty structs in C as a compiler extension and such a struct has size 0.
This answer is wrong. In addition to the size thing, the constructor/destructor part is also inaccurate: structs can have constructors and destructors just fine. In fact, struct creates full-fledged classes in C++.
U
UKMonkey

While implied by other answers, it's not explicitly mentioned - that structs are C compatible, depending on usage; classes are not.

This means if you're writing a header that you want to be C compatible then you've no option other than struct (which in the C world can't have functions; but can have function pointers).


M
Maciej Urbański

There exists also unwritten rule that tells: If data members of class have no association with itself, use struct. If value of data member depends on another value of data member, use class.

f.e

class Time
{
    int minutes;
    int seconds;
}

struct Sizes
{
    int length;
    int width;
};

C
Community

You might consider this for guidelines on when to go for struct or class, https://msdn.microsoft.com/en-us/library/ms229017%28v=vs.110%29.aspx .

√ CONSIDER defining a struct instead of a class if instances of the type are small and commonly short-lived or are commonly embedded in other objects. X AVOID defining a struct unless the type has all of the following characteristics: It logically represents a single value, similar to primitive types (int, double, etc.). It has an instance size under 16 bytes. It is immutable. It will not have to be boxed frequently.


That's about .NET/C#, not C++.
s
snow

Class is only meaningful in the context of software engineering. In the context of data structures and algorithms, class and struct are not that different. There's no any rule restricted that class's member must be referenced.

When developing large project with tons of people without class, you may finally get complicated coupled code because everybody use whatever functions and data they want. class provides permission controls and inherents to enhance decoupling and reusing codes.

If you read some software engineering principles, you'll find most standards can not be implemented easily without class. for example: http://en.wikipedia.org/wiki/SOLID_%28object-oriented_design%29

BTW, When a struct allocates a crunch of memory and includes several variables, value type variables indicates that values are embbeded in where struct is allocated. In contrast, reference type variable's values are external and reference by a pointer which is also embedded in where struct is allocated.


This answer doesn't seem to be about C++. In C++, struct declares classes (including permission controls, inheritance, etc).
M
MOBITI MAHENGE

The difference between struct and class keywords in C++ is that, when there is no specific specifier on particular composite data type then by default struct or union is the public keywords that merely considers data hiding but class is the private keyword that considers the hiding of program codes or data. Always some programmers use struct for data and class for code sake. For more information contact other sources.


OP already mentioned that struct members are public by default, and the accepted answer from 2009 mentions inheritance. Why write another answer that adds no new information 4 years later?
W
Wageesha

Out of all these factors,it can be concluded that concept Class is highly suitable to represent real world objects rather than "Structures".Largely because OOP concepts used in class are highly practical in explaining real world scenarios therefore easier to merge them to reality.For an example,default inheritance is public for structs but if we apply this rule for real world,it's ridiculous.But in a class default inheritance is private which is more realistic.

Anyways,what i need to justify is Class is a much broader,real world applicable concept whereas Structure is a primitive Concept with poor internal organization(Eventhough struct follows OOP concepts,they have a poor meaning)


How is private inheritance "more realistic"? Most programming languages don't even have the concept of private inheritance.
T
Tushar Bobade

The main difference between structure and class keyword in oops is that, no public and private member declaration present in structure.and the data member and member function can be defined as public, private as well as protected.


This answer doesn't seem to be about C++. In C++, struct declares classes (including public/private/protected, inheritance, etc).
Y
Yug Rawal

The main difference between struct and class is that in struct you can only declare data variables of different data types while in class you can declare data variables,member functions and thus you can manipulate data variables through functions.

-> another handy thing that i find in class vs struct is that while implementing files in a program if you want to make some operations of a struct again and again on every new set of operations you need to make a separate function and you need to pass object of struct after reading it from the file so as to make some operations on it . while in class if you make a function that does some operations on the data needed everytime..its easy you just have to read object from file and call the function..

But it depennds on the programmer which way he/she finds suitable...according to me i prefer class everytime just because it supports OOPs and thats the reason it is implemented in almost every languages and its the wonderful feature of all time programming ;-)

And yeah the most unforgotten difference i forgot to mention is that class supports data hiding and also supports operations that are performed on built in data types while struct doesnt !


In C++ struct can have member functions
@chtz Thanks... i got to know about this feature just now.. thanks a lot.. i m just a beginer not as expert as you so in future also need advices from u guys :-)
and actually, member function is to be intended in its wider meaning. struct can have constructors, member function and distructor (possibly virtual, possibly private, protected or public). They can inherit and be inherited. so actually... everything a class can have. a tipical convention is to declare with struct what you would have declared as struct in C, and as class if we declare private and protected variable, virtual functions, ctor and dtor, and so on. But it's just a convention, not enforced by the language
A
Ali

I found an other difference. if you do not define a constructor in a class, the compiler will define one. but in a struct if you do not define a constructor, the compiler do not define a constructor too. so in some cases that we really do not need a constructor, struct is a better choice (performance tip). and sorry for my bad English.


This is not true. There is no difference between a class defined with the struct keyword and a class defined with the class keyword in this respect.
when i use a struct with out a constructor, i get a better speed than a class with out a constructor.
@Ali No you don't.
F
Frank Myat Thu

Classes are Reference types and Structures are Values types. When I say Classes are reference types, basically they will contain the address of an instance variables.

For example:

Class MyClass
{
    Public Int DataMember;  //By default, accessibility of class data members 
                            //will be private. So I am making it as Public which 
                            //can be accessed outside of the class.
}

In main method, I can create an instance of this class using new operator that allocates memory for this class and stores the base address of that into MyClass type variable(_myClassObject2).

Static Public void Main (string[] arg)
{
    MyClass _myClassObject1 = new MyClass();
    _myClassObject1.DataMember = 10;

    MyClass _myClassObject2 = _myClassObject1;
    _myClassObject2.DataMember=20;
}

In the above program, MyClass _myClassObject2 = _myClassObject1; instruction indicates that both variables of type MyClass

myClassObject1 myClassObject2

and will point to the same memory location. It basically assigns the same memory location into another variable of same type.

So if any changes that we make in any one of the objects type MyClass will have an effect on another since both are pointing to the same memory location.

"_myClassObject1.DataMember = 10;" at this line both the object’s data members will contain the value of 10. "_myClassObject2.DataMember = 20;" at this line both the object’s data member will contains the value of 20. Eventually, we are accessing datamembers of an object through pointers.

Unlike classes, structures are value types. For example:

Structure MyStructure
{
    Public Int DataMember;  //By default, accessibility of Structure data 
                            //members will be private. So I am making it as 
                            //Public which can be accessed out side of the structure.
}

Static Public void Main (string[] arg)
{
    MyStructure _myStructObject1 = new MyStructure();
    _myStructObject1.DataMember = 10;

    MyStructure _myStructObject2 = _myStructObject1;
    _myStructObject2.DataMember = 20;
}

In the above program, instantiating the object of MyStructure type using new operator and storing address into _myStructObject variable of type MyStructure and assigning value 10 to data member of the structure using "_myStructObject1.DataMember = 10".

In the next line, I am declaring another variable _myStructObject2 of type MyStructure and assigning _myStructObject1 into that. Here .NET C# compiler creates another copy of _myStructureObject1 object and assigns that memory location into MyStructure variable _myStructObject2.

So whatever change we make on _myStructObject1 will never have an effect on another variable _myStructObject2 of type MyStructrue. That’s why we are saying Structures are value types.

So the immediate Base class for class is Object and immediate Base class for Structure is ValueType which inherits from Object. Classes will support an Inheritance whereas Structures won’t.

How are we saying that? And what is the reason behind that? The answer is Classes.

It can be abstract, sealed, static, and partial and can’t be Private, Protected and protected internal.


I think this is trying to answer the question for c#? The question is about c++, and that makes this answer inaccurate, right?
@smw absolutely, he has written his sample code in C# and I came here to check that there wasn't a difference in parameter passing between C++ struct and class. No one here has mentioned it, nor the entire web. So I don't think this applies to C++ either.
This isn't even valid C# code - the keywords are all wrong. "Public", "Static", "Int", and "Class" should not be capitalized and it should be "struct" not "Structure". Description is mostly accurate for C# (though structs can still inherit interfaces, just not other structs), but doesn't answer the question since it was about C++...
This answer is complete nonsense. It uses a strange mix of Visual Basic and C# for its code examples, and the whole "Classes are Reference types and Structures are Values types"... thing only applies to .NET (ie C#) not to C++. You're in the wrong room, dude. This is where you belong : stackoverflow.com/questions/13049
But this answer indicates the same is valid for C(++), that structs are used like value types. int a = 3; int b = a; and similar MyStruct a=MyStruct("data"); MyStruct b = a;
N
Naveen

There are 3 basic difference between structure and class

1St- memory are reserved for structure in stack memory (which is near to programming language )whether for class in stack memory are reserved for only reffrence and actual memory are reserved in heap memory.

2Nd - By default structure treat as a public whether class treat as a private .

3Rd- can't re -use code in structure but in class we can re-use same code in many time called inhertence


C++ structures do support inheritance and abstract methods. Structure and class are allocated the same way. Allocating class on the heap even if they are not instantiated with a "new" doesn't make sense for embedded systems.
#1 is also wrong. The stack is used when you declare something of that type as a local variable, and the heap is used when you use the "new" operator.