Reading floats from a binary file

[Matt McGonigle]

Hi all,

Please help me out with this. Perhaps it is a dumb
question, but I can't seem to make it work. I am doing a
file conversion using an unformatted binary file for input
and outputting to a normal text file. I need to read in a
float from the binary file, but it sets my input stream to
failbit. Is there a special way I can read in floats? All
I am doing is

ifstream inFile("input.d ta");
float limitValue;

inFile >> limitValue;

to input it. Can someone please tell me how to go about
getting this to work?

Thanks,
Matt

[Eric Twietmeyer]

Re: Reading floats from a binary file

"David Lowndes" <davidl@mvps.or g> wrote in message
news:busaiv02sg t0388negvve8soi fk1a3krja@4ax.c om...[color=blue][color=green]
> >Please help me out with this. Perhaps it is a dumb
> >question, but I can't seem to make it work. I am doing a
> >file conversion using an unformatted binary file for input
> >and outputting to a normal text file. I need to read in a
> >float from the binary file, but it sets my input stream to
> >failbit. Is there a special way I can read in floats? All
> >I am doing is
> >
> >ifstream inFile("input.d ta");
> >float limitValue;
> >
> >inFile >> limitValue;[/color]
>
> Matt,
>
> What's the format of your file? Is it just a stream of single
> precision (32-bit) IEEE floating point values - or something else?
>
> Dave
> --
> MVP VC++ FAQ: http://www.mvps.org/vcfaq[/color]

When using iostream input operator you need to be reading in text versions
of the values. If your file consists of space separated items of the form
"3.14159", etc, then you can use input operator >> to read into a float
variable.

However, you indicate you have a binary file, so presumably your format is
in IEEE (or something, it doesn't matter), but it is not a text
representation. In this case you will need to do "raw" reads from the
stream using the methods "read" or "readsome" of basic_istream. You then
convert (depending on the format of the data in the file) to an IEEE float
and store in your float variable (generally using appropriate casting
operations).

HTH,
Eric Twietmeyer

[David Lowndes]

Re: Reading floats from a binary file

>The file is mixed - there are character values, integers,[color=blue]
>and float values. It is the output of a test machine, and
>has all kinds of information in it, only one subset of
>which is a float.[/color]

OK, so are you doing the right thing with your code?

Eric is right (I'd missed that aspect), the ifstream operators you're
using will do text conversions rather than direct reads. For example
the following shows writing and reading a binary floating point value:

{
ofstream oFile("input.dt a");
float fVal = 123.456f;

oFile.write( (const char *)&fVal, sizeof( fVal ) );
}
{
ifstream inFile("input.d ta");
float limitValue;

inFile.read( (char*)&limitVa lue, sizeof( limitValue ) );
limitValue = limitValue;
}

Dave
--
MVP VC++ FAQ: http://www.mvps.org/vcfaq