Read C structure in java

[Colin Bester]

I am needing to read a binary file from disk which is created with a 'C' application. All contents of the file are floats.
Is there any conversion routine to enable me to read these four bytes into a java float?
(Or suggestions)
Thanks,
Colin

[Michael Morris]

Hi Colin,
What C method are you using to dump the floats to the file and what OS/hardware is the file created on? If you're sure of the byte order then you should be able the use the static method float Float.intBitsToFloat(int bits) to reconstruct them for you. This method uses the IEEE 754 floating-point "single format" bit layout. You'll obviously have to create the int with the four bytes before passing it in but that shouldn't be a big deal.
To read in the file, I would suggest using a FileInputStream and calling the read(byte[] buf) method where buf is a 4 member array. Just setup a loop like:

    try {
        FileInputStream fis = new FileInputStream("floatfile");
        byte[] buf = new byte[4];
        while (fis.read(buf) != -1) {
            /* Process buf and create float */
        }
    } catch (IOException ioe) {
        /* Deal with exception */
    }

Hope this helps,
Michael Morris

[Colin Bester]

Thanks appreciate the input. I must have looked a dozen times at the api's and didn't 'see' the int to float function. All working smoothly now.
Colin

[Gopi Balaji]

Colin,
Out of curiosity, what is the byte order of data in the file? Does that machine have an Intel/DEC chip, or any other?
-GB.

[Colin Bester]

The data is created on an Intel machine where the lsb is stored first. i.e. 0x01020304 is store in memory as 04,03,02,01 (Little Endian). In my code I swopped these around and made a new number which I passed to IntBitsToFloat to get my float value.
Trust this helps.
Colin

[Gopi Balaji]

Originally posted by Colin Bester:
The data is created on an Intel machine where the lsb is stored first. i.e. 0x01020304 is store in memory as 04,03,02,01 (Little Endian). In my code I swopped these around and made a new number which I passed to IntBitsToFloat to get my float value.
Trust this helps.
Colin

Thanks.
-GB.