Problems with reading binary data files

[Dimitri Papoutsis]

Hi there, ...

i want to read in a binary data file and to cast each pair of 2 bytes to a
signed short value. i try to different ways, but none results in what i
suppose it should result.

#include <stdio.h>
#include <stdlib.h>
#include <iostream.h>
#include <fstream.h>

int main(){
Â*Â*Â*Â*Â*Â*Â*Â *charÂ*file_nam e[13];
Â*Â*Â*Â*Â*Â*Â*Â *FILE*Â*fptr;
Â*Â*Â*Â*Â*Â*Â* *signedÂ*short *c;
Â*Â*Â*Â*Â*Â*Â*Â *
Â*Â*Â*Â*Â*Â*Â* *printf("Enter *aÂ*filename:Â* ");
Â*Â*Â*Â*Â*Â*Â*Â *gets(file_name );
Â*Â*Â*Â*Â*Â*Â*Â *
Â*Â*Â*Â*Â*Â*Â*Â *//Â*VariationÂ*A: Â*UsingÂ*theÂ*s tandartÂ*fopen, Â*fgetc,Â*fclos eÂ*functions
Â*Â*Â*Â*Â*Â*Â*Â *ifÂ*((fptrÂ*=Â *fopen(file_nam e,"rb"))Â*==Â*0 ){
Â*Â*Â*Â*Â*Â*Â* *Â*Â*Â*Â*Â*Â*Â* Â*printf("Sorry .Â*Can'tÂ*open *fileÂ*\"%s\"\n ",Â*file_na me);
Â*Â*Â*Â*Â*Â*Â*Â *Â*Â*Â*Â*Â*Â*Â* Â*exitÂ*(EXIT_F AILURE);
Â*Â*Â*Â*Â*Â*Â*Â *}
Â*Â*Â*Â*Â*Â*Â*Â *for(intÂ*i=0;i <11;i++){
Â*Â*Â*Â*Â*Â*Â*Â *Â*Â*Â*Â*Â*Â*Â* Â*cÂ*=Â*fgetc(f ptr);
Â*Â*Â*Â*Â*Â*Â*Â *Â*Â*Â*Â*Â*Â*Â* Â*coutÂ*<<Â*cÂ* <<Â*"\t";
Â*Â*Â*Â*Â*Â*Â*Â *}
Â*Â*Â*Â*Â*Â*Â*Â *coutÂ*<<Â*"\n" ;
Â*Â*Â*Â*Â*Â*Â*Â *fclose(fptr);
Â*Â*Â*Â*Â*Â*Â*Â *//Â*VariationÂ*B: Â*UsingÂ*theÂ*f streamÂ*class
Â*Â*Â*Â*Â*Â*Â*Â *charÂ*buffer[24];
Â*Â*Â*Â*Â*Â*Â*Â *ifstreamÂ*date i(file_name,Â*i os::inÂ*|Â*ios: :binary);Â*Â*Â* Â*Â*Â*Â*
Â*Â*Â*Â*Â*Â*Â*Â *if(!datei){
Â*Â*Â*Â*Â*Â*Â* *Â*Â*Â*Â*Â*Â*Â* Â*printf("Sorry .Â*Can'tÂ*open *fileÂ*\"%s\"\n ",Â*file_na me);
Â*Â*Â*Â*Â*Â*Â*Â *Â*Â*Â*Â*Â*Â*Â* Â*exitÂ*(EXIT_F AILURE);
Â*Â*Â*Â*Â*Â*Â*Â *}
Â*Â*Â*Â*Â*Â*Â*Â *datei.read(buf fer,Â*24);
Â*Â*Â*Â*Â*Â*Â*Â *forÂ*(intÂ*i=0 ;i<24;i++){
Â*Â*Â*Â*Â*Â*Â*Â *Â*Â*Â*Â*Â*Â*Â* Â*coutÂ*<<Â*(si gnedÂ*short)buf fer[i]Â*Â*<<Â*"\t";
Â*Â*Â*Â*Â*Â*Â*Â *}
Â*Â*Â*Â*Â*Â*Â*Â *coutÂ*<<Â*"\n" ;

Â*Â*Â*Â*Â*Â*Â*Â *returnÂ*EXIT_S UCCESS;
}

You can test the program with the following data file:
Â*Â*Â*Â*Â*Â*Â*Â *http://www.uni-koblenz.de/~bastard/dat.1

The result that i suppose my program to print out can be generated with the
following perl-script:
Â*Â*Â*Â*Â*Â*Â*Â *http://www.uni-koblenz.de/~bastard/binaryDataRead

=> ./binaryDataRead dat.1
=> Result :
0Â*Â*Â*Â*Â*Â*Â*-489Â*Â*Â*Â*-458Â*Â*Â*Â*31Â* Â*Â*Â*Â*Â*474Â* Â*Â*Â*Â*-260Â*Â*Â*Â*-214Â*Â*Â*Â*-88Â*Â*Â*Â*Â*-241Â*Â*Â*Â*-112Â*Â*Â*
212Â*Â*Â*Â*Â*39 3Â*Â*Â*Â*Â*390

whereas the "0" is just an index of the printed record.

My Program results in:
=> Variation A:
23Â*Â*Â*Â*Â*Â*2 54Â*Â*Â*Â*Â*54Â *Â*Â*Â*Â*Â*254Â *Â*Â*Â*Â*31Â*Â* Â*Â*Â*Â*0Â*Â*Â* Â*Â*Â*Â*218Â*Â* Â*Â*Â*1Â*Â*Â*Â* Â*Â*Â*252Â*Â*Â* Â*Â*254Â*Â*Â*Â*
42Â*Â*Â*Â*Â*Â*2 55

=> Variation B:
23Â*Â*Â*Â*Â*Â*-2Â*Â*Â*Â*Â*Â*54 Â*Â*Â*Â*Â*Â*-2Â*Â*Â*Â*Â*Â*31 Â*Â*Â*Â*Â*Â*0Â* Â*Â*Â*Â*Â*Â*-38Â*Â*Â*Â*Â*1Â* Â*Â*Â*Â*Â*Â*-4Â*Â*Â*Â*Â*Â*-2Â*Â*Â*Â*Â*
42Â*Â*Â*Â*Â*Â*-1Â*Â*Â*Â*Â*Â*-88Â*Â*Â*Â*Â*-1Â*Â*Â*Â*Â*Â*15 Â*Â*Â*Â*Â*Â*-1Â*-112Â*Â*Â*Â*Â*-1Â*Â*Â*Â*Â*Â*-44Â*Â*Â*Â*Â*0Â* Â*Â*Â*Â*Â*
-119Â*Â*Â*Â*1Â*Â *Â*Â*Â*Â*Â*-122Â*Â*Â*Â*1

About Variation A:
i completely misunderstand the causal dependencies between the code and the
result.

About Variation B:
i see that i always cast and print out just 1 Byte. How can i cast and print
out 2 Bytes ? Are my approaches completly wrong ?

i hope anybody can help...

bye
Dimi

[Scott McPhillips [MVP]]

Dimitri Papoutsis wrote:

Hi there, ...

i want to read in a binary data file and to cast each pair of 2 bytes to a
signed short value. i try to different ways, but none results in what i
suppose it should result.
datei.read(buff er, 24);
for (int i=0;i<24;i++){
cout << (signed short)buffer[i] << "\t";
}

buffer[i] reads a char prior to the cast. To read 2 bytes you can
dereference a pointer to a 2 byte type.

signed short* p = (signed short*) &buffer[i];
signed short result = *p;

You will also need to modify your loop so i increments 0, 2, 4, ...

--
Scott McPhillips [VC++ MVP]

[Thomas Matthews]

Scott McPhillips [MVP] wrote:

Dimitri Papoutsis wrote:

Hi there, ...

i want to read in a binary data file and to cast each pair of 2 bytes
to a
signed short value. i try to different ways, but none results in what i
suppose it should result.
datei.read(buff er, 24);
for (int i=0;i<24;i++){
cout << (signed short)buffer[i] << "\t";
}

buffer[i] reads a char prior to the cast. To read 2 bytes you can
dereference a pointer to a 2 byte type.

signed short* p = (signed short*) &buffer[i];
signed short result = *p;

You will also need to modify your loop so i increments 0, 2, 4, ...

A better method for reading multibyte binary integers is to
read each byte individually and build the value. This will
allow the proper reading of the data file regardless of the
Endianism of the platform.

--
Thomas

[red floyd]

Dimitri Papoutsis wrote:

Hi there, ...

i want to read in a binary data file and to cast each pair of 2 bytes to a
signed short value. i try to different ways, but none results in what i
suppose it should result.

#include <stdio.h>
#include <stdlib.h>
#include <iostream.h> non standard header. Use <iostream>
#include <fstream.h> non standard header. Use <fstream>


int main(){

[redacted]

[Old Wolf]

Scott McPhillips [MVP] wrote:

Dimitri Papoutsis wrote:

Hi there, ...

i want to read in a binary data file and to cast each pair of 2 bytes to a signed short value. i try to different ways, but none results in what i suppose it should result.
datei.read(buff er, 24);
for (int i=0;i<24;i++){
cout << (signed short)buffer[i] << "\t";
}
buffer[i] reads a char prior to the cast. To read 2 bytes you can
dereference a pointer to a 2 byte type.

signed short* p = (signed short*) &buffer[i];

Undefined behaviour, if buffer + i isn't correctly aligned
for a short. Also, 'short' is signed, you don't have to write
out the word signed.

Better would be:

short result;
memcpy(&result, buffer, sizeof result);

although this may not work on esoteric systems.

You will also need to modify your loop so i increments 0, 2, 4, ...

--
Scott McPhillips [VC++ MVP]