在C中声明数组大小

Question

Its quite embarrassing but I really want to know... So I needed to make a conversion program that converts decimal(base 10) to binary and hex. I used arrays to store values and everything worked out fine, but i declared the array as int arr[1000]; because i thought 1000 was just an ok number, not too big, not to small...someone in class said " why would you declare an array of 1000? Integers are 32 bits". I was too embarrased to ask what that meant so i didnt say anything. But does this mean that i can just declare the array as int arr[32]; instead? Im using C btw

Answer 1

No, the int type has tipically a 32 bit size, but when you declare

int arr[1000];

you are reserving space for 1000 integers, ie 32'000 bits, while with

int arr[32];

you can store up to 32 integers.

You are practically asking yourself a question like this: if an apple weighs 32 grams, I want to my bag to contain 1000 apples or 32 apples?

Answer 2

Don't be embarrassed. Fear is your enemy and in the end you will be perceived based on contexts that you have no hope of significantly influencing. Anyway, to answer your question, your approach is incorrect. You should declare the array with a size completely determined by the number of positions used.

Concretely, if you access the array at 87 distinct positions (from 0 to 86) then you need a size of 87.

Answer 3

0 to 4,294,967,295 is the maximum possible range of numbers you can store in 32 bits.If your number is outside this range you cannot store your number in 32 bits.Since each bit will occupy one index location of your array if you number falls in that range array size of 32 will do fine.for example consider number 9 it will be stored in array as a[]={1,0,0,1} .

In order to know the know range of numbers, your formula is 0 to (2^n -1) where n is the number of bits in binary. means in the array size of 4 or 4 bits you can just store number from range 0 to 15 .

In C , integer datatype can store typically up to 2,147,483,647 and 4,294,967,295 if you are using unsigned integer. Since the maximum value, an integer data type can store in C is within the range of maximum possible number which can be expressed using 32 bits. It is safe to say that array size of 32 is the best size for defining an array.Sice you will never require more than 32 bits to express any number using an int.

Answer 4

I will use

int a = 42;
char bin[sizeof a * CHAR_BIT + 1];
char hex[sizeof a * CHAR_BIT / 4 + 1]

I think this include all possibility.

Answer 5

Consider that also the 'int' type is ambiguous. Generally it depends on the machine you're working on and at minimum its ranges are: -32767,+32767:

https://en.wikipedia.org/wiki/C_data_types

Can I suggest to use the stdint types?

int32_t/uint32_t

Answer 6

What you did is okay. If that is precisely what you want to do. C is a language that lets you do whatever you want. Whenever you want. The reason you were berated on the declaration is because of 'hogging' memory. The thought being, how DARE YOU take up space that is possibly never used... it is inefficient.

And it is. But who cares if you just want to run a program that has a simple purpose? A 1000 16 or 32 bit block of memory is weeeeeensy teeeeny tiny compared to computers from the way back times when it was necessary to watch over how much RAM you were taking up. So - go ahead.

But what they should have said next is how to avoid that. More on that at the end - but first a thing about built in data types in C.

An int can be 16 or 32 bits depending on how you declare it. And your compiler's settings...

A LONG int is 32.

consider:

short int x = 10;    // declares an integer that is 16 bits 
signed int x = 10;   // 32 bit integer with negative and positive range
unsigned int x = 10  // same 32 bit integer - but only 0 to positive values

To specifically code a 32 bit integer you declare it 'long'

long int = 10;           // 32 bit
unsigned long int = 10;  // 32 bit 0 to positive values

Typical nomenclature is to call a 16 bit value a WORD and a 32 bit value a DWORD - (double word). But why would you want to type in:

long int x = 10;

instead of:

int x = 10;

?? For a few reasons. Some compilers may handle the int as a 16 bit WORD if keeping up with older standards. But the only real reason is to maintain a convention of strongly typecasted code. Make it read directly what you intend it to do. This also helps in readability. You will KNOW when you see it = what size it is for sure, and be reminded whilst coding. Many many code mishaps happen for lack of attention to code practices and naming things well. Save yourself hours of headache later on by learning good habits now. Create YOUR OWN style of coding. Take a look at other styles just to get an idea on what the industry may expect. But in the end you will find you own way in it.

On to the array issue ---> So, I expect you know that the array takes up memory right when the program runs. Right then, wham - the RAM for that array is set aside just for your program. It is locked out from use by any other resource, service, etc the operating system is handling.

But wouldn't it be neat if you could just use the memory you needed when you wanted, and then let it go when done? Inside the program - as it runs. So when your program first started, the array (so to speak) would be zero. And when you needed a 'slot' in the array, you could just add one.... use it, and then let it go - or add another - or ten more... etc.

That is called dynamic memory allocation. And it requires the use of a data type that you may not have encountered yet. Look up "Pointers in C" to get an intro.

If you are coding in regular C there are a few functions that assist in performing dynamic allocation of memory:

 malloc and free      ~ in the alloc.h library routines

in C++ they are implemented differently. Look for:

 new and delete

A common construct for handling dynamic 'arrays' is called a "linked-list." Look that up too...

Don't let someone get your flustered with code concepts. Next time just say your program is designed to handle exactly what you have intended. That usually stops the discussion.

• Atomkey