C# How are arrays stored in memory

Question

I guess my main question is, will this always work as long as I don't re-initialize (new byte[#]) the array that was passed as the parameter?

static unsafe decimal GetDecimal(byte[] ba)
{
    decimal* decimal_PTR;
    fixed (byte* byte_PTR = &ba[0])
    {
        decimal_PTR = ((decimal*)byte_PTR);
    }
    return *decimal_PTR;
}

I'm not sure how C# handles arrays in memory. I didn't even know they were managed types until about an hour ago. I just want to know if I pass in a decimal as a byte[], will it always return the correct value? Any other information you can provide is appreciated.

Also, Would it be better to create a decimal at the beginning of the function, then in the fixed block, make it equal the value pointed to by ((decimal*)byte_PTR) then return the decimal that was created? — MJLaukala
– MJLaukala, Commented Jun 27, 2012 at 4:58
May I know how you obtain the byte[]?. Answer would depend on this information. — Amit Mittal
– Amit Mittal, Commented Jun 27, 2012 at 11:35
the byte[] is part of a message sent over a network connection using NetworkStream.Read(byte[] buffer, int offset, int size). The example method above is slightly different then the implementation but the array passed to the method is the same array that was passed to NetworkStream.Read. This is a windows only program so I do believe that I do not need to worry about endianness. — MJLaukala
– MJLaukala, Commented Jun 28, 2012 at 9:44
I guess my question is will the array of bytes always be stored in order right next to each other in memory or is it possible for the bytes to be stored in separate "chunks" of memory. I want to make sure that part of the array can't be stored somewhere else causing my decimal pointer to return an incorrect value. — MJLaukala
– MJLaukala, Commented Jun 28, 2012 at 9:45
I guess an example of what I am trying to say is: Will each byte of byte[5] be stored as such. Address 00000000 = byte[0] Address 00000001 = byte[1] Address 00000002 = byte[2] Address 00000003 = byte[3] Address 00000004 = byte[4] Or can it be: Address 00000000 = byte[0] Address 00000001 = byte[1] Address 00000005 = byte[2] Address 00000007 = byte[3] Address 00000008 = byte[4] My main goal is to retrieve a decimal from a byte[], which BitConverter does not contain a method for this. — MJLaukala
– MJLaukala, Commented Jun 28, 2012 at 9:54

Amit Mittal · Accepted Answer · 2012-06-28 10:52:26Z

3

After @MJLaukala's clarification through comments:

An Array in .Net is a block of contiguous memory just like any other language. But unlike C/C++ this block not only stores the array's elements but also other 'information' including rank and length of the array. Having said this, &ba[0] will return the pointer to the first element of the array and you can safely increment the pointer to go till the last element. ba.Length will provide the number of elements in the array.
Now if you are sure that the byte array represents the decimal exactly then your code will work fine.

Memory layout of decimal (128 bits):

First 2 bytes are 0

3rd byte contain a value between 0 and 28, indicating the power of 10 to divide the 96-bit integer part by to produce the Decimal value

4th byte : first 7 bits are zero, 8th bit indicates the sign of the decimal (1 meaning negative)

Next 12 bytes: Integer part of the Decimal.

edited Jun 28, 2012 at 10:52

answered Jun 27, 2012 at 11:33

Amit Mittal

2,75619 silver badges25 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

MJLaukala Over a year ago

Thank you very much for the clarification! This is info i definitely will not forget.

Cole Tobin Over a year ago

Why couldn't it just be struct Decimal { long integer; long fractional; } and is represented by integer.fractional? 42.13 would be { integer = 42; fractional = 13 }.

Amit Mittal Over a year ago

This is a classic 'why is floating-point representation preferred over fixed-point representation' question. Your notation is a fixed point notation with max value of the order of 10^18 (check max value of long). Floating-point representations (such as decimal's) allow greater range. Max value of decimal is of the order of 10^29.

Collectives™ on Stack Overflow

C# How are arrays stored in memory

1 Answer 1

3 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related