[Date Index] [Thread Index] [Author Index]

Float32 from a file

• To: mathgroup at smc.vnet.net
• Subject: [mg5826] Float32 from a file
• From: rudy at msunews.cl.msu.edu (Rudolf B. Blazek)
• Date: Thu, 23 Jan 1997 03:19:30 -0500
• Organization: Michigan State University
• Sender: owner-wri-mathgroup at wolfram.com

Hi Mma group.

I have to read a binary file with float32 numbers in it. The file was most 
likely created with a C program.

1) I am about to write a Mathematica function that will read 4 bytes at a 
time and compute what the real number was. 

2) I am also considering writing a little C program to read the dataset and 
return the real numbers as ASCII. Then I would call the utility from 
Mathematica. 

Could someone tell me if they've done something like that?

I like the solution 1) for being written completely in Mma. But it will be 
slow.

I like the solution 2) for being fast and using some C built in mechanism for 
writing/reading such files with automatic conversion of the 4 bytes into a 
real number. But one needs to compile the C program for each platform, put it 
into some reasonable place, call it from Mma... :-(((

Questions:

a) Would you know how the conversion of the 4 bytes into the real number 
should look like? I found some reference to a booklet with an ANSI standard 
for the float32 real numbers, but I don't have the booklet.

b) If you wrote a utility similar to what I described. Would you share it?

c) Could you possibly give me a short example how float32 type is read from 
and external file in C, C++, or Objective C ? 

Thank you, a nice day to all.

Rudy

--
Rudy Blazek                        Michigan State University
blazek at stt.msu.edu                 Department of Statistics & Probability