Thread: need help storing input values into an array

Code:

if (customKey == '0'){ lonarray[0] == '0'; i++;}
if (customKey == '1'){ lonarray[i] == '1'; i++;}
if (customKey == '2'){ lonarray[1] == '2'; i++;}
if (customKey == '3'){ lonarray[i] == '3'; i++;}
if (customKey == '4'){ lonarray[i] == '4'; i++;}
if (customKey == '5'){ lonarray[i] == '5'; i++;}
if (customKey == '6'){ lonarray[i] == '6'; i++;}
if (customKey == '7'){ lonarray[i] == '7'; i++;}
if (customKey == '8'){ lonarray[i] == '8'; i++;}
if (customKey == '9'){ lonarray[i] == '9'; i++;}
if (customKey == '.'){ lonarray[i] == '.'; i++;}
if (customKey == '-'){ lonarray[i] == '-'; i++;}

You probably want those to be assignments (single '=').

Thanks, I knew it was something simple, is there an easier more sound way of turning an array into an integer without using the "atoi" function? I would like it perform some sort of mathematics with the entered keys.
I should imagine if its possible, it would be tricky due to a potential character being a '-' or a '.'.

Thanks again.

Are the expected inputs simple integers, or real numbers? This will determine the reasonable options to consider.

If you are able to use library functions (if they are supported in your environment, and you have enough space to do so), this would probably be best. "strtol", "strtod", and "sscanf" are some good options.

I can tell that you're working on an arduino chip. That means you have some limitations.

One idea to work with decimal points without a special type would be to read values like this: sscanf (lonarray, "%d.%d", &x, &y);

Then you can use a sliding scale (the sign variable is to handle negative values properly):

sign = x < 0;
x *= -1;
x = x * 10 + y;
if (sign) x += -1;

y = 2 * x;

You would have to take care to display the answer correctly though. For example if you had access to printf it would work like this:
printf ("%d.%d", y / 10, y % 10);
Negatives are slightly different:
if (sign) y *= -1;
printf ("%c%d.%d", sign? '-' : ' ', y/10, y%10);

I hope you can figure it out from here.

Hi, thanks for the advice, I'll give it a go, could you please explain why the function "atof" appears to work fine, I had a look around and the only disadvantage I could find was to do with strange problems occurring if certain characters were included.
If this is indeed the only limitation, atof should be sufficient as only the specified characters from the if statements will ever be included in the array to begin with.
I do intend on performing trigonometric functions with the variables, but I assume that won't be an issue.

I know I have to use floats for decimal numbers but if the answer is a whole number, e.g. 1, the answer still prints 1.00, not that its a problem, but would be nice if could save space, bearing this in mind, the answerers that i have acquired so far have indeed been correct, however they only seem to display to 2 decimal places which is insufficient(I have a feeling though, if I increase the decimal places, whole numbers will print lik "1.00000000" etc.

Anyway back to your post, how would the program respond (using both your method and atof) if for example, a "-" is pressed at any point other to the beginning, or the "." is detected more than once in the array?

And to answer your printing query: printf could just be replaced by Serial.print and as for lcd.print i think its just a case of quotes for text then a comma then the variable, but to be honest i think its clearer just to have two lcd.print statements one for the text and one for the variable.

Thanks again

Something like this?
if something is wrong is just displays "0.00..."
is that correct?
on another note, do you know why by default without the 5 for decimal places, it is by default 2 decimal places and always ends in "00"
for example if I do -1.5x2 (without the 5 in lcd print) it returns "-3.00"
if I do -1.028*2(with the 5 dp's) it returns -2.056"00"
it would make more sense if it just displayed -2.056, only go to 5 decimal places as long as the value continues to be not 0.

Code:

if (customKey == 'B')  {
    float x;
    float y;
    lcd.clear();  
    lcd.print("nmbrs prsd were:");
    lcd.setCursor(0,1);
    x=strtod(lonarray, NULL);
     y=(x*-2);
    lcd.print(y,5); //5 decimal places
    Serial.print(y);
  }

Something like this?
if something is wrong is just displays "0.00..."
is that correct?

No, calling strtod with a NULL is not any better than atof(). You should be doing something more like this:

Code:

#include <errno.h>
char *end;
errno = 0;
x = strtod(lonarray, &end);
if (errno || end == lonarray || *end != '\0') {
   lcd.print("Error");
   /* reset the input loop, abort, whatever's appropriate */
}
y = -2 * x;

What you do with y and x mathematically changes between examples so I hope I got it right.

on another note, do you know why by default without the 5 for decimal places, it is by default 2 decimal places and always ends in "00"
for example if I do -1.5x2 (without the 5 in lcd print) it returns "-3.00"
if I do -1.028*2(with the 5 dp's) it returns -2.056"00"
it would make more sense if it just displayed -2.056, only go to 5 decimal places as long as the value continues to be not 0.

I don't think you can print y directly. At least according to reference, you should not; float is not among the supported types for data. You should use code to convert y to a separate string variable. There you will have an opportunity to format your answer to N decimal places.

hi, is the following what is intended:

if I enter "9.8.44"
when I multiply by two it prints""error-19.6"
another example:
"-3.8-5"
prints "Error 7.60"

It is roughly what is intended. You are responsible for getting the code to work exactly how you want.