Thread: Help discarding a space in char array

I have some embedded sql code that reads data from a sql database and inserts the data into a ingres isql database. That works fine. My problem lies in the code when I use atof to convert a character field (set up by outside company) in sql to a double for the ingres database. This value should have been an integer in SQL as well. So I store this char field from sql into a variable called rslt_rec_char.Hardness and if there is a space before the number (seems to be only problem) I need to remove it so the atof function will work. Can someone help me make changes to the code below to do this? Thanks. The int is_double just looks for decimal point and the function is posted below also.

Code:

    879         if (is_double(rslt_rec_char.Hardness))
    880             rslt_rec.Hardness = atof(rslt_rec_char.Hardness);
    881         else if (!strcmp(rslt_rec_char.Hardness,"")) 
    882             rslt_rec.Hardness = 0.0;
    883         else
    884         {
    885             rslt_rec.Hardness = 0.0;
    886             sprintf(jnl.mesg, "ERROR: CIN %s. Hardness test value invalid - \"%s\".\n"
    887                 "Hardness in test record %s not recorded!",
    888                 rslt_rec_char.CIN, rslt_rec_char.Hardness, rslt_rec_char.ResultsID);
    889             log_it(logfile,jnl);
    890             send_email(db_nm,"Met lab test upload error",jnl.mesg);
    891         }

Code:

   1421 int
   1422 is_double(char s[])
   1423 {
   1424     int index, decimal_point_found = 0;
   1425
   1426     if (!strcmp(s,"")) return 0; 
   1427
   1428     for (NULL; *s && isspace(*s); s++);
   1429     if (s[0] == '-') s++;
   1430
   1431     for (index = 0; index < strlen(s); index++)
   1432     {
   1433         if (s[index] == '.') decimal_point_found++;
   1434         else if (!isdigit(s[index])) return 0;
   1435     }
   1436
   1437     return ((decimal_point_found <= 1) ? 1 : 0);
   1438 }

Code:

char * first_nonblank = rslt_rec_char.Hardness ; 
while (*first_nonblank) { 
   if ( isspace(*first_nonblank) ) first_nonblank++ ; 
   else break ; 
}

then use first_nonblank inside the atof() function.

Surely this:

Code:

     for (NULL; *s && isspace(*s); s++);

is much clearer as:

Code:

     while(*s && isspace(*s))
          s++;

If nothing else, you can certainly remove the NULL before the first semicolon.

Also, you isdigit doesn't allow the

Code:

E{,+,-}[0-9]*

format of floating point.

--
Mats

Yes, you have to print the string pointed to by first_nonblank to see the space removed. It really not physically "removed", it is just not part of the new string pointed to by first_nonblank, which is what you need for atof().

hardness2 is a double, not a string.

Personally, I find this:

Code:

while (*first_nonblank)
{
if (isspace(*first_nonblank))
first_nonblank++;
else break;
}

easier to read and understand as:

Code:

while (*first_nonblank && isspace(*first_nonblank))
{
    first_nonblank++;
}

Also, since the null character is not a space, you could simplify that to:

Code:

while (isspace(*first_nonblank))
{
    first_nonblank++;
}

which is even more straightforward to read and understand.

Good simplification laserlight.

Todd

Originally Posted by cjohnman

I plugged your simplified version of Todds code in and it worked great! Thanks for your feedback. If there was more than one space this would work and no worrys about memory leak right?

There are no memory allocations in the code posted here, so I don't see how it could leak anything.

--
Mats

My idea would be to add an additional variable, which is originally zero, then increment it within the loop, and if it's greater than one when the loop finishes, you've got an error.

--
Mats

Actually, in that case I do not think you need a loop. If the first two characters are spaces, then there is an error. Else, if only the first character is a space, then you ignore it. Else, you can convert immediately.