By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
424,987 Members | 1,552 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 424,987 IT Pros & Developers. It's quick & easy.

beginner fscanf question

P: n/a
Hi

I have a beginner question concerning fscanf.
First I had a text file which just contained some
hex numbers:

0C100012
0C100012
....

It was easy to read this values into an integer variable:
FILE *fp = fopen("test.dat","r");
int mem_word = 0;
while (fscanf(fp,"%x", &mem_word) != EOF)
....

Now, I have the following format where my hex number
is splitted by spaces. So I wonder if there is
an easy way to parse such a line and not doing
a lot of space removing etc.

0C 10 00 12 // 00000000
0C 10 00 12 // 00000000
....

Thanks for helpful comments.
Aug 15 '07 #1
Share this Question
Share on Google+
10 Replies


P: n/a
Roman Zeilinger wrote:
Now, I have the following format where my hex number
is splitted by spaces. So I wonder if there is
an easy way to parse such a line and not doing
a lot of space removing etc.
I don't think so. But it's trivial to remove the spaces and then use
sscanf().
0C 10 00 12 // 00000000
0C 10 00 12 // 00000000
Or you could write your own simple parser
Aug 15 '07 #2

P: n/a
Roman Zeilinger <Ro*************@yahoo.dewrote:
Hi
I have a beginner question concerning fscanf.
First I had a text file which just contained some
hex numbers:
0C100012
0C100012
...
It was easy to read this values into an integer variable:
FILE *fp = fopen("test.dat","r");
int mem_word = 0;
while (fscanf(fp,"%x", &mem_word) != EOF)
...
Now, I have the following format where my hex number
is splitted by spaces. So I wonder if there is
an easy way to parse such a line and not doing
a lot of space removing etc.
0C 10 00 12 // 00000000
0C 10 00 12 // 00000000
If there's always " // 00000000" at the end of the line and
never anything else (except trailing white space) it's rela-
tively simple:

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

int main( void ) {
FILE *fp = fopen( "sh.dat", "r" );
unsigned int d[ 4 ];
unsigned int x;

if ( fp == NULL )
return EXIT_FAILURE;

while ( fscanf( fp, "%x%x%x%x // 00000000",
d, d + 1, d + 2, d + 3 ) == 4 ) {
x = ( ( ( ( ( d[ 0 ] << 8 ) + d[ 1 ] ) << 8 )
+ d[ 2 ] ) << 8 ) + d[ 3 ];
printf( "x = %x\n", x );
}

return 0;
}

but if that's not the case then you will need to do a bit more of
work in that while loop:

while ( fscanf( fp, "%x%x%x%x", d, d + 1, d + 2, d + 3 ) == 4 ) {
x = ( ( ( ( ( d[ 0 ] << 8 ) + d[ 1 ] ) << 8 )
+ d[ 2 ] ) << 8 ) + d[ 3 ];
printf( "x = %x\n", x );
while ( ( c = fgetc( fp ) ) != '\n' && c != EOF )
/* empty */ ;
}

You also will need to add a definition of 'c' (as an int, not a
char!).

You could use something simpler if you can be sure that there will
be never more than 12 characters following your quadruplet of hex
numbers (again except trailing white space characters):

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

int main( void ) {
FILE *fp = fopen( "sh.dat", "r" );
unsigned int d[ 4 ];
unsigned int x;
char buffer[ 13 ];

if ( fp == NULL )
return EXIT_FAILURE;

while ( fscanf( fp, "%x%x%x%x%12[^\n]",
d, d + 1, d + 2, d + 3, buffer ) == 5 ) {
x = ( ( ( ( ( d[ 0 ] << 8 ) + d[ 1 ] ) << 8 )
+ d[ 2 ] ) << 8 ) + d[ 3 ];
printf( "x = %x\n", x );
}

return 0;
}
Regards, Jens
--
\ Jens Thoms Toerring ___ jt@toerring.de
\__________________________ http://toerring.de
Aug 15 '07 #3

P: n/a
You could use something simpler if you can be sure that there will
be never more than 12 characters following your quadruplet of hex
numbers (again except trailing white space characters):

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

int main( void ) {
FILE *fp = fopen( "sh.dat", "r" );
unsigned int d[ 4 ];
unsigned int x;
char buffer[ 13 ];

if ( fp == NULL )
return EXIT_FAILURE;

while ( fscanf( fp, "%x%x%x%x%12[^\n]",
d, d + 1, d + 2, d + 3, buffer ) == 5 ) {
x = ( ( ( ( ( d[ 0 ] << 8 ) + d[ 1 ] ) << 8 )
+ d[ 2 ] ) << 8 ) + d[ 3 ];
printf( "x = %x\n", x );
}

return 0;
}
Great thanks Jens. It works fine for one line so I have to figure out
how I can parse the whole txt file :)

Cheers!
Aug 15 '07 #4

P: n/a
I modified the code from jens slightly, but
it looks like the program sticks with the
first line it reads from "icache.dat". Why
does the program not move to the next line of
the text file after the fscanf and remains with the
first one?

int main( void ) {
FILE *fp = fopen( "icache.dat", "r" );
unsigned int d[ 4 ];
unsigned int x;

if ( fp == NULL )
return EXIT_FAILURE;

while ( fscanf( fp, "%x%x%x%x%", d, d + 1, d + 2, d + 3 ) != EOF )
{
x = ( ( ( ( ( d[ 0 ] << 8 ) + d[ 1 ] ) << 8 )
+ d[ 2 ] ) << 8 ) + d[ 3 ];
printf( "x = %x\n", x );
}
return 0;
}

Cheers!
Aug 15 '07 #5

P: n/a
Roman Zeilinger wrote:
I modified the code from jens slightly, but
it looks like the program sticks with the
first line it reads from "icache.dat". Why
does the program not move to the next line of
the text file after the fscanf and remains with the
first one?
Time for the C FAQ, I suspect. http://c-faq.com

There's loads of discussion of the pitfalls and snares of scanf..
Aug 15 '07 #6

P: n/a
Roman Zeilinger <Ro*************@yahoo.dewrote:
I modified the code from jens slightly, but
it looks like the program sticks with the
first line it reads from "icache.dat". Why
does the program not move to the next line of
the text file after the fscanf and remains with the
first one?
int main( void ) {
FILE *fp = fopen( "icache.dat", "r" );
unsigned int d[ 4 ];
unsigned int x;
if ( fp == NULL )
return EXIT_FAILURE;
while ( fscanf( fp, "%x%x%x%x%", d, d + 1, d + 2, d + 3 ) != EOF )
{
x = ( ( ( ( ( d[ 0 ] << 8 ) + d[ 1 ] ) << 8 )
+ d[ 2 ] ) << 8 ) + d[ 3 ];
printf( "x = %x\n", x );
}
return 0;
}
If there are some unusable data in that line as in your origi-
nal example had you don't skip them. fscanf() will try to read
them also as 4 hex numbers, fail for obvious reasons, return 0
and not move forward in the file. And you only test the return
value for EOF and repeat if you get anything else. That makes
fscanf() look again at the stuff it already deemed as not being
4 hex numbers the last time round, giving you the same result
again etc., so you end up in an endless loop. That's why I had
proposed to either read in everything of the remainder of the
line with a loop over getc() until you find the newline charac-
terer or by using a long enough dummy buffer that gets filled
with the unusable data in the call of fscanf().

Regards, Jens
--
\ Jens Thoms Toerring ___ jt@toerring.de
\__________________________ http://toerring.de
Aug 15 '07 #7

P: n/a
jt@toerring.de (Jens Thoms Toerring) writes:
Roman Zeilinger <Ro*************@yahoo.dewrote:
>I modified the code from jens slightly, but
it looks like the program sticks with the
first line it reads from "icache.dat". Why
does the program not move to the next line of
the text file after the fscanf and remains with the
first one?
>int main( void ) {
FILE *fp = fopen( "icache.dat", "r" );
unsigned int d[ 4 ];
unsigned int x;
> if ( fp == NULL )
return EXIT_FAILURE;
> while ( fscanf( fp, "%x%x%x%x%", d, d + 1, d + 2, d + 3 ) != EOF )
{
x = ( ( ( ( ( d[ 0 ] << 8 ) + d[ 1 ] ) << 8 )
+ d[ 2 ] ) << 8 ) + d[ 3 ];
printf( "x = %x\n", x );
}
return 0;
}

If there are some unusable data in that line as in your origi-
nal example had you don't skip them. fscanf() will try to read
them also as 4 hex numbers, fail for obvious reasons, return 0
and not move forward in the file.
For this reason it is almost always better to test the return value
from fscanf for success (== 4 in this case) rather than for one
particular failure (from the possible EOF, 0, 1, 2 or 3).

--
Ben.
Aug 15 '07 #8

P: n/a
Roman Zeilinger wrote:
>
Hi

I have a beginner question concerning fscanf.
First I had a text file which just contained some
hex numbers:

0C100012
0C100012
...

It was easy to read this values into an integer variable:

FILE *fp = fopen("test.dat","r");
int mem_word = 0;
while (fscanf(fp,"%x", &mem_word) != EOF)
...

Now, I have the following format where my hex number
is splitted by spaces. So I wonder if there is
an easy way to parse such a line and not doing
a lot of space removing etc.

0C 10 00 12 // 00000000
0C 10 00 12 // 00000000
...

Thanks for helpful comments.
I like to read lines and proccess strings.

/* BEGIN beginner.c */

#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <ctype.h>
#include <assert.h>

#define FILE_NAME ("test.dat")
#define LINE ("0C 10 00 12")
#define LINELENGTH 11
#define str(x) # x
#define xstr(x) str(x)

void hex_squeeze(char *s1);

int main(void)
{
unsigned long mem_word;
char array[sizeof LINE];
int rc;
char *endptr;
char *fn = FILE_NAME;
FILE *fp = fopen(fn, "r");

assert(LINELENGTH == sizeof LINE - 1);
fp = fopen(fn, "r");
if (fp == NULL) {
printf("fopen problem with %s\n", fn);
exit(EXIT_FAILURE);
}
do {
rc = fscanf(fp, "%" xstr(LINELENGTH) "[^\n]%*[^\n]", array);
if (!feof(fp)) {
getc(fp);
}
if (rc == 0) {
array[0] = '\0';
}
if (rc == 1) {
hex_squeeze(array);
mem_word = strtoul(array, &endptr, 16);
printf("%lx\n", mem_word);
}
} while (rc != EOF);
fclose(fp);
return 0;
}

void hex_squeeze(char *s)
{
char *p = s;

do {
if (isxdigit((unsigned char)*s)) {
*p++ = *s;
}
} while (*s++ != '\0');
*p = '\0';
}

/* END beginner.c */
--
pete
Aug 16 '07 #9

P: n/a
In article <5i*************@mid.uni-berlin.de>
Jens Thoms Toerring <jt@toerring.dewrote:

[a bunch of correct stuff with one subtle flaw]
unsigned int d[ 4 ];
unsigned int x;
[snippage]
x = ( ( ( ( ( d[ 0 ] << 8 ) + d[ 1 ] ) << 8 )
+ d[ 2 ] ) << 8 ) + d[ 3 ];
printf( "x = %x\n", x );
To make this portable, you need to give "x" the type "unsigned
long", and convert at least one of the various operands in the
shift and add sequence to "unsigned long". Of course then the
printf() format argument must be "%lx".

(An implementation with 16-bit "unsigned int"s will turn:

x = (((((d[0] << 8) + d[1]) << 8) + d[2]) << 8) + d[3];

into the equivalent of just:

x = (d[2] << 8) + d[3];

since the extra bits will "fall off the end". A lot of
implementations today have 32-bit "int", where the problem
will not occur; and some of 64-bit "long", where "unsigned
long" is perhaps overkill, but the oberkill is portable.)
--
In-Real-Life: Chris Torek, Wind River Systems
Salt Lake City, UT, USA (4039.22'N, 11150.29'W) +1 801 277 2603
email: forget about it http://web.torek.net/torek/index.html
Reading email is like searching for food in the garbage, thanks to spammers.
Aug 16 '07 #10

P: n/a
Chris Torek <no****@torek.netwrote:
In article <5i*************@mid.uni-berlin.de>
Jens Thoms Toerring <jt@toerring.dewrote:
unsigned int d[ 4 ];
unsigned int x;
[snippage]
x = ( ( ( ( ( d[ 0 ] << 8 ) + d[ 1 ] ) << 8 )
+ d[ 2 ] ) << 8 ) + d[ 3 ];
printf( "x = %x\n", x );
To make this portable, you need to give "x" the type "unsigned
long", and convert at least one of the various operands in the
shift and add sequence to "unsigned long". Of course then the
printf() format argument must be "%lx".
(An implementation with 16-bit "unsigned int"s will turn:
x = (((((d[0] << 8) + d[1]) << 8) + d[2]) << 8) + d[3];
into the equivalent of just:
x = (d[2] << 8) + d[3];
since the extra bits will "fall off the end". A lot of
implementations today have 32-bit "int", where the problem
will not occur; and some of 64-bit "long", where "unsigned
long" is perhaps overkill, but the oberkill is portable.)
Thank you, I should have realized that myself;-)

Regards, Jens
--
\ Jens Thoms Toerring ___ jt@toerring.de
\__________________________ http://toerring.de
Aug 16 '07 #11

This discussion thread is closed

Replies have been disabled for this discussion.