fscanf

[Bill Cunningham]

I'm doing something wrong and all I know to do is turn to clc. I have a
text file containing 2 doubles separated by a tab.

..26 0

Is the text. I want to read the two double and printf them out. Here's my
file.

#include <stdio.h>

int main() {
FILE *fp;
double x,y;
fp=fopen("zo"," r"); /*error checking out for brevity */
fscanf(fp,"%.2f \t%.2f",&string );
fclose(fp);
printf("%.2f%.2 f",x,y);
}

All I get is garbage that is contained in x and y. For whatever simple
reason that is beyond me evidently I can't read and printf out to stdin from
this text file. I don't think fread is really necessary.

Bill

[Bill Cunningham]

"Bill Cunningham" <no****@nspam.c omwrote in message
news:zulqk.194$ p72.57@trnddc05 ...

Ok I'll try it. I'm not used to using the *scanf family. I usually use
fgets or fgetc.

Here's the code re-written a little but the results are the same.

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

int main()
{
FILE *fp;
double x, y;
int rv;
char *string = "%.2f\t%.2f ";
if ((fp = fopen("zo", "r")) == NULL) {
puts("open error");
exit(-1);
}
rv = fscanf(fp, string, &x, &y);
fclose(fp);
printf("%i\n", rv);
printf(string, x, y);
}

The the rv that catches the return value of fscanf reported 0. The same
garbage was printed from the x and y variables.

0.00 4.87

Bill

[Bill Cunningham]

"Keith Thompson" <ks***@mib.orgw rote in message
news:ln******** ****@nuthaus.mi b.org...

vi******@gmail. com writes:

>On Aug 18, 5:23 am, Keith Thompson <[e-mail] rote:

>correction (typo): cppreference.co m not cVppreference.c om.

Whoops, thanks for the correction.

Keith I hope this posts with no header problem above. I have k&r2. I use
the www.cppreference.com C section to look up return values mainly. I find
linux's man pages very vague IMO on my linux anyway, that is old.

Bill

[Ben Bacarisse]

"Bill Cunningham" <no****@nspam.c omwrites:

"Bill Cunningham" <no****@nspam.c omwrote in message
news:zulqk.194$ p72.57@trnddc05 ...

>Ok I'll try it. I'm not used to using the *scanf family. I usually use
fgets or fgetc.

Here's the code re-written a little but the results are the same.

<snip>

Compare with:

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

int main(void)
{
FILE *fp;
double x, y;
int rv;
if ((fp = fopen("zo", "r")) == NULL) {
perror("open error");
return EXIT_FAILURE;
}
rv = fscanf(fp, "%lf\t%lf", &x, &y);
fclose(fp);
printf("%i\n", rv);
printf("%.2f\t% .2f\n", x, y);
return EXIT_SUCCESS;
}

fscanf formats are not the same as printf formats.

--
Ben.

[Richard Heathfield]

Bill Cunningham said:

I have k&r2.

Great.

I use the www.cppreference.com C section to look up return values mainly.

I suggest you stop doing that. You have a much more authoritative reference
available to you: K&R2. Use it.

--
Richard Heathfield <http://www.cpax.org.uk >
Email: -http://www. +rjh@
Google users: <http://www.cpax.org.uk/prg/writings/googly.php>
"Usenet is a strange place" - dmr 29 July 1999

[Bill Cunningham]

"Bill Cunningham" <no****@nspam.c omwrote in message
news:yYoqk.212$ lf2.25@trnddc07 ...

??

Who are you addressing Barry?

Bill

If you are making a comment to simply say something I'm not listening.
If you want to help and are saying something critical I'm listening.

There are 2 main people on usenet. The one who asks for help and those
who want to help. Then there's noise. Meaningless, helpless, nonsense.

Bill

[Bill Cunningham]

"Ben Bacarisse" <be********@bsb .me.ukwrote in message
news:87******** ****@bsb.me.uk. ..

fscanf formats are not the same as printf formats.

I hadn't noticed that Ben for some reason. So what I should do is use
%lf or the float type for x and y then.

Now I feel silly for asking a question like this but I was really stuck.
But isn't %f still a float conversion in *scanf ? The program I wrote to
print the file "zo" and the text it contains uses fprintf %f conversion for
2 doubles. Here's a snippet...

double x,y;
x=strtod(argv[1],NULL);
y=strtod(argv[2],NULL);
fprintf(argv[3],"%.2f\t%.2f",x ,y);

Float conversion can be used on doubles.
But I guess not in fscanf.

Bill

[Paul]

Bill Cunningham wrote:

"Bill Cunningham" <no****@nspam.c omwrote in message
news:zulqk.194$ p72.57@trnddc05 ...

>Ok I'll try it. I'm not used to using the *scanf family. I usually use
fgets or fgetc.

Here's the code re-written a little but the results are the same.

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

int main()
{
FILE *fp;
double x, y;
int rv;
char *string = "%.2f\t%.2f ";
if ((fp = fopen("zo", "r")) == NULL) {
puts("open error");
exit(-1);
}
rv = fscanf(fp, string, &x, &y);
fclose(fp);
printf("%i\n", rv);
printf(string, x, y);
}

The the rv that catches the return value of fscanf reported 0. The same
garbage was printed from the x and y variables.

0.00 4.87

Bill

So now you know that zero parameters were converted by fscanf.
X and Y cannot have good values, because no conversion was done.

This guy wrote a similar example, but even he did not consider that
fscanf might convert zero items. He is checking for EOF with
feof(file). If you had continued to use the "append" mode for
your input file, then inserting an feof(file) check might
have indicated a problem (i.e. already at the end of the file).

(second example down...)
http://irc.essex.ac.uk/www.iota-six....nf_fprintf.asp

Since you changed to "r" mode for opening the file, then the
file pointer is probably not at the end of the file any more.

And that leaves the "string" specification as being the dodgy bit.
Try some other specifications first. Try converting just one
number with "%f" and work from there. This stuff looks tricky.

See PDF page 299, for a few examples of fscanf usage. Or search for
the string "25 54.32E-1 thompson" which is input to the first example.

http://www.open-std.org/jtc1/sc22/wg...docs/n1124.pdf

I was amazed at all the stuff I could find.

Paul

[Ben Bacarisse]

"Bill Cunningham" <no****@nspam.c omwrites:

"Ben Bacarisse" <be********@bsb .me.ukwrote in message
news:87******** ****@bsb.me.uk. ..

>fscanf formats are not the same as printf formats.

I hadn't noticed that Ben for some reason. So what I should do is use
%lf or the float type for x and y then.

No. You really do need a reference manual. I used lf because you
don't have floats. For scanf, %lf expects a pointer to a double (as
you have).

Now I feel silly for asking a question like this but I was really stuck.
But isn't %f still a float conversion in *scanf ?

Yes it is. It would work if you had float variables but you don't,
you have doubles.

The program I wrote to
print the file "zo" and the text it contains uses fprintf %f conversion for
2 doubles. Here's a snippet...

double x,y;
x=strtod(argv[1],NULL);
y=strtod(argv[2],NULL);
fprintf(argv[3],"%.2f\t%.2f",x ,y);

Float conversion can be used on doubles.

By "float conversion" do you mean %f? If so, then it *requires* a
double in the printf. The trick is that floats get converted to
double in a call to printf anyway so there is little danger of getting
it wrong.

--
Ben.

[Keith Thompson]

"Bill Cunningham" <no****@nspam.c omwrites:

"Keith Thompson" <ks***@mib.orgw rote in message
news:ln******** ****@nuthaus.mi b.org...

>BTW, do you expect us to *guess* what's in your "zo" file?

No Keith I've already said it contains,

.26 0.00

So you did.

However, you only mentioned that in your original article, in which
you posted non-compilable code. I didn't take the time to go back and
re-read it.

[snip]

--
Keith Thompson (The_Other_Keit h) ks***@mib.org <http://www.ghoti.net/~kst>
Nokia
"We must do something. This is something. Therefore, we must do this."
-- Antony Jay and Jonathan Lynn, "Yes Minister"

[Keith Thompson]

"Bill Cunningham" <no****@nspam.c omwrites:

"Keith Thompson" <ks***@mib.orgw rote in message
news:ln******** ****@nuthaus.mi b.org...

>vi******@gmail. com writes:

>>On Aug 18, 5:23 am, Keith Thompson <[e-mail] rote:

>>correction (typo): cppreference.co m not cVppreference.c om.

Whoops, thanks for the correction.

Keith I hope this posts with no header problem above. I have
k&r2. I use the www.cppreference.com C section to look up return
values mainly. I find linux's man pages very vague IMO on my linux
anyway, that is old.

As I recall, K&R2's reference section does a perfectly good job of
describing return types for standard library functions.

I reiterate my advice: Stop using any reference materials other than
K&R2. Well, not quite; the comp.lang.c FAQ is also an excellent and
reliable resource. If those two resources together are not
sufficiently helpful, then ask here -- but *only* after you've
carefully checked the types and meanings of any arguments and return
types for all standard functions you're calling.

--
Keith Thompson (The_Other_Keit h) ks***@mib.org <http://www.ghoti.net/~kst>
Nokia
"We must do something. This is something. Therefore, we must do this."
-- Antony Jay and Jonathan Lynn, "Yes Minister"