Thread: sscanf problem...help please!

Originally Posted by victor_miami

The "value" variable is a double (please see the code below). Is there a way to assign this to an array of float type variable? (i.e. - var_value[10][1600], where 10 can be up to 10 different variables and the 1600 can be 1600 pieces of data per variable).

Short answer: yes.

Originally Posted by victor_miami

The reason I need to do this is that I am writing this out to a file and do not want it to appear as double precision. Or am I better off handling it when I write it to the file (at least I think there is a way to do this but I need to first figure out how).

I think now I'd take this discussion in a different direction.

What do you want your file format to be?
Do you intend to have a text file or a binary file?
Would an array of structures be more appropriate?
How do you intend to keep track of the actual number of values?
Etc.?

Originally Posted by victor_miami

The file will be a text file. I don't know if it will make a difference if the values are written out as float or char type? Does it make any difference? Another process will eventually open this text file and read the values from it and use them.

I currently use a structure with the variable name and their respective values.

I keep track of the number of values through a loop for variables, so each variable that comes up in the loop then has all of it's values loaded into a value array. I'm certain there are far better ways of doing this but this is the tac I've taken (right or wrong).

Your approach sounds quite reasonable. My simulation developed into something like this.

Code:

struct record
{
   char   name[128];
   double value[1600];
   size_t length;
};

/* ... */

               if ( end != field && *end == '\0' )
               {
                  /* It's a number. */
                  rec[j].value[rec[j].length++] = value;
               }
               else
               {
                  /* It's not a number. */
                  j = count++;
                  strcpy(rec[j].name, field);
                  i = 0;
               }

So it sounds quite similar to what you've done.

I don't think using a float instead of a double will have much benefit for you. I was using %g in the fprintf to get rid of trailing zeros anyways.

You mean like casting ?

Edit : Whoa...I was ridiculously beat on this thread