Excuse me for my trivial question, but I'd like to know how to convert byte array to float? What I could think of are cast(float)(byte[]) and to!float(byte[]) but they do not work for me. Thanks!

On Friday, 17 July 2015 at 18:43:27 UTC, DLangLearner wrote:
> Excuse me for my trivial question, but I'd like to know how to convert byte array to float? What I could think of are cast(float)(byte[]) and to!float(byte[]) but they do not work for me. Thanks!

You want to include [] in the cast so:



byte[] b = [1, 2, 3, 4]
byte[] f = cast(float[])b;

writeln(b);
writeln(f);

You can even cast on a slice:

byte[] b = [1, 2, 3, 4, 5, 6, 7, 8]
byte[] f = cast(float[])b[4..8];

On Friday, 17 July 2015 at 18:53:24 UTC, byron wrote:
> On Friday, 17 July 2015 at 18:43:27 UTC, DLangLearner wrote:
>> Excuse me for my trivial question, but I'd like to know how to convert byte array to float? What I could think of are cast(float)(byte[]) and to!float(byte[]) but they do not work for me. Thanks!
>
> You want to include [] in the cast so:
>
>
>
> byte[] b = [1, 2, 3, 4]
> byte[] f = cast(float[])b;
>
> writeln(b);
> writeln(f);
>
> You can even cast on a slice:
>
> byte[] b = [1, 2, 3, 4, 5, 6, 7, 8]
> byte[] f = cast(float[])b[4..8];

Ah I miss read, if you want as a float, not a float array you can do:

byte[] b = [1, 2, 3, 4];
float f = *cast(float*)b.ptr;

not sure if there is a better way

On 2015-07-17 20:58, byron wrote:

> Ah I miss read, if you want as a float, not a float array you can do:
>
> byte[] b = [1, 2, 3, 4];
> float f = *cast(float*)b.ptr;
>
> not sure if there is a better way

I think a union can be used as well, not sure if it's better though.

-- 
/Jacob Carlborg

On Friday, 17 July 2015 at 19:03:46 UTC, Jacob Carlborg wrote:
> On 2015-07-17 20:58, byron wrote:
>
>> Ah I miss read, if you want as a float, not a float array you can do:
>>
>> byte[] b = [1, 2, 3, 4];
>> float f = *cast(float*)b.ptr;
>>
>> not sure if there is a better way
>
> I think a union can be used as well, not sure if it's better though.

not a bad idea.

testing on asm.dlang.org both reduce to the same assembly
(-release -inline -O)

float pointerTest(inout byte[4] b) {
   return *cast(float*)b.ptr;
}

float unionTest(inout byte[4] b) {
  union U {
   byte[4] b;
    float f;
  }
  return U(b).f;
}

On Friday, 17 July 2015 at 18:58:33 UTC, byron wrote:
> On Friday, 17 July 2015 at 18:53:24 UTC, byron wrote:
>> On Friday, 17 July 2015 at 18:43:27 UTC, DLangLearner wrote:
>>> Excuse me for my trivial question, but I'd like to know how to convert byte array to float? What I could think of are cast(float)(byte[]) and to!float(byte[]) but they do not work for me. Thanks!
>>
>> You want to include [] in the cast so:
>>
>>
>>
>> byte[] b = [1, 2, 3, 4]
>> byte[] f = cast(float[])b;
>>
>> writeln(b);
>> writeln(f);
>>
>> You can even cast on a slice:
>>
>> byte[] b = [1, 2, 3, 4, 5, 6, 7, 8]
>> byte[] f = cast(float[])b[4..8];
>
> Ah I miss read, if you want as a float, not a float array you can do:
>
> byte[] b = [1, 2, 3, 4];
> float f = *cast(float*)b.ptr;
>
> not sure if there is a better way

thanks for quick reply, i'll plug above lines into my codes.

On Friday, July 17, 2015 18:43:26 DLangLearner via Digitalmars-d-learn wrote:
> Excuse me for my trivial question, but I'd like to know how to
> convert byte array to float? What I could think of are
> cast(float)(byte[]) and to!float(byte[]) but they do not work for
> me. Thanks!

You could use std.bitmanip.peek or std.bitmanip.read - though you do need to make sure that you're telling it the right endianness, depending on how the array of bytes was generated.

- Jonathan M Davis