Link to home
Start Free TrialLog in
Avatar of Meir Rivkin
Meir RivkinFlag for Israel

asked on

file exception

hi experts,

i need to read 2 columns of foating points feagures from a text files, something like this:

2.3424   564.685
3.5465   45.567
and so on...

there is at least one space between 2 feagures and the floating point number is unknown which means that it might be one number after floating point and might be 4 numbers after floating point.

i'm using this code right now:

     ifstream in(strFileName.GetBuffer(0));

     double mass, score;
     CArray<double,double> arrMass, arrScore;
     
     while(!in.eof())
     {
          in >> mass >> score;
          arrMass.Add(mass);
          arrScore.Add(score);
     }

the problem is that i wanna use file exception in order to catch reading exceptions, such as string instead of floating points feagures, or more than 2 floating points feagures in one row.

what is the best way to achieve it?
Avatar of Meir Rivkin
Meir Rivkin
Flag of Israel image

ASKER

in case of courrupted file the 'while(!in.eof()' cause endless loop...
         
Avatar of curri
curri

You can use in.fail() instead of in.eof()

in.fail() returns true if the operation failed, for any reason.

Orlando
By default stream objects do not throw exceptiosn on error conditions.  However, the exeptions() member funciton can be used to set flags for the types of error conditions on which exceptions should be thrown.  for example, when a formatted extraction fails due to invalid input, it is considered a "fail" error.  This usualy results only in the fail bit being set, but if the stream is set  to throw exceptions on this error, an exception will also be thrown.

ifstream in(strFileName.GetBuffer(0));

double mass, score;
CArray<double,double> arrMass, arrScore;

in.exceptins(ios_base::failbit)
                   

try
{
   while(!in.eof())
   {
      in >> mass >> score;
      arrMass.Add(mass);
      arrScore.Add(score);
   }
}
catch (const ios_base::failure &f)
{
}
nietod: then why it keeps looping on an on?
besides i used your code and it never reached to the 'catch' point.

the endless loop takes place when the file looks like this:
2.45 34.676
345 6234rtbe rtb
reg.rger 546g erb54

i want the loop to stop if he can't assign readen text from the file to the floating points values...
it should work.  It works for me in my tests.

Did you remember to set the exceptiojn flags with the exceptions() function?

Post the exact code you used.
i copyied your code as is, but it kept running.
curri: thanks for your comment, i managed to handle the reading properly but i realy like to know how to determine the reason of the failure?

the fail bit is set only from one cause.  When a formatted read encounters input that is invalid for the data type to be read.  For example, if you try to read an integer and float and the first character read is not a digit.

This will set the failbit flag and depending on your current exceptions settings will throw an exception.

Why didn't you post your code?
what is this line? - in.exceptins(ios_base::failbit)
                 

Hi sedgwick:

If fail() and not eof(), then you know that something strange happened. Most probably bad inputs (hardware errors are very rare).

I can't find any way of getting more info about the error (do you know any nietod?) and it kinda makes sense. It seems to be hard to find errors that apply to any kind of file (hard disks, removable media, network filesystems), and not worth the trouble.

Of course, you can always read character by character and form the floats yourself, but that's probably more trouble that you want :)

I think also, if you are reading from an fstream, there may be a way of getting to the underlying C or even OS file, and you can get info from there.

Orlando
ASKER CERTIFIED SOLUTION
Avatar of nietod
nietod

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
thanks, workd just fine