How to split a floating number into 4 individual digits??? (Arduino)

[grimreeper132]

How do I split a floating number up into 4 digits, so I can use each one individually so I can display using LEDs. The number I am getting are 25.12 for example which would relate to 2 Red LEDs, 5 Yellow, 1 white and 2 Blue. I can do the part after getting the numbers into the 4 digits split up. The only way I can think of its through a massively long bit of code of ifs going through all possibilities which just seams stupid and that there is a better way to do it. Any Suggestions welcome.

[WereCatf]

Is the value always two integers and two decimals, like the "25.12" you mentioned? Or in other words, is the value always less than 100 and more or equal to 0?

[tikker]

Are they always four digits? Then a modulo trick will work:

At the risk of solving homework  here it is. First multiply by 100, cast to int and then loop over it taking modulus 10 to give you the last digit of the integer.

x = int(x * 100)
while x > 10:
  digit = x % 10
  x = x / 10

 

[Mira Yurizaki]

Multiply it by 100 to shift the digits. You can then truncate it to an int and do your magic from there.

[grimreeper132]

1 hour ago, WereCatf said:

Is the value always two integers and two decimals, like the "25.12" you mentioned? Or in other words, is the value always less than 100 and more or equal to 0?

No negatives between 0 and 100 (not including 100)

 

1 hour ago, tikker said:

Are they always four digits? Then a modulo trick will work:

At the risk of solving homework  here it is. First multiply by 100, cast to int and then loop over it taking modulus 10 to give you the last digit of the integer.

x = int(x * 100)
while x > 10:
  digit = x % 10
  x = x / 10

 

It's not homework it's me being a nerd in my own time deciding to do something stupidly pointless cause I can. Also that's not arduino code.

 

2 hours ago, M.Yurizaki said:

Multiply it by 100 to shift the digits. You can then truncate it to an int and do your magic from there.

Yea, I did that, thanks, and it worked. It took me far too long though to work out that I was getting the digits in reverse order (at least 30 mins, me smurt)

[grimreeper132]

2 hours ago, M.Yurizaki said:

Multiply it by 100 to shift the digits. You can then truncate it to an int and do your magic from there.

on another note what does this error message mean???

 

Quote

Arduino: 1.8.5 (Windows 10), Board: "Arduino Due (Programming Port)"

In file included from C:\Users\Grimreeper132\AppData\Local\Arduino15\packages\arduino\hardware\sam\1.6.11\cores\arduino/Arduino.h:34:0,

                 from sketch\Version_1.ino.cpp:1:

C:\Users\Grimreeper132\AppData\Local\Arduino15\packages\arduino\hardware\sam\1.6.11\cores\arduino/binary.h:31:12: error: expected unqualified-id before numeric constant

 #define B1 1

            ^

D:\Code\Temperture\Temperture and variable Resistor to LED Output\Version_1\Version_1.ino:48:11: note: in expansion of macro 'B1'

 const int B1=34;

           ^

exit status 1
Error compiling for board Arduino Due (Programming Port).

This report would have more information with
"Show verbose output during compilation"
option enabled in File -> Preferences.
 

 

[Mira Yurizaki]

3 minutes ago, grimreeper132 said:

on another note what does this error message mean???

It's complaining you're trying to make 1 = 34.

 

When you have a #define preprocessor, the compiler looks for all instances of the first argument and replaces it with the second argument. In the case of #define B1 1, anywhere in the code where there's a B1, the compiler replaces it with 1.

[JacobFW]

8 hours ago, grimreeper132 said:

How do I split a floating number up into 4 digits, so I can use each one individually so I can display using LEDs. The number I am getting are 25.12 for example which would relate to 2 Red LEDs, 5 Yellow, 1 white and 2 Blue. I can do the part after getting the numbers into the 4 digits split up. The only way I can think of its through a massively long bit of code of ifs going through all possibilities which just seams stupid and that there is a better way to do it. Any Suggestions welcome.

I'd really recommend rethinking your encoding mechanism for this.  Floating point numbers cannot accurately represent all numbers in their ranges, and using them in calculations can introduce additional error:

 

i.e.  (0.12 - 0.02) is not guaranteed to equal 0.1

and 0.1 cannot be represented accurately by an floating point number.

[Unimportant]

12 hours ago, grimreeper132 said:

How do I split a floating number up into 4 digits, so I can use each one individually so I can display using LEDs. The number I am getting are 25.12 for example which would relate to 2 Red LEDs, 5 Yellow, 1 white and 2 Blue. I can do the part after getting the numbers into the 4 digits split up. The only way I can think of its through a massively long bit of code of ifs going through all possibilities which just seams stupid and that there is a better way to do it. Any Suggestions welcome.

Note that on such small microcontrollers, such as AVR used in arduino, floating point is expensive in both program memory used and execution time. Floating point is complex and without any hardware support it has to be completely emulated on these small 8-bit controllers.

 

A common trick to reduce the overhead is to stick to integers but simply multiply by a power of 10 first. For example, say your result of 25.12 is the result of dividing 628 by 25. This could be implemented with integers as 62800 / 25 = 2512. If you need to print this result on a display, you simply print a decimal point in right place. In your case you can simply divide by 1000 to get the red leds, mod 1000 divide by 100 for the yellow leds, and so on.

 

Even if this requires using 32 bit integers it will still be much cheaper then using float.

[tikker]

11 hours ago, grimreeper132 said:

No negatives between 0 and 100 (not including 100)

 

It's not homework it's me being a nerd in my own time deciding to do something stupidly pointless cause I can. Also that's not arduino code.

 

Yea, I did that, thanks, and it worked. It took me far too long though to work out that I was getting the digits in reverse order (at least 30 mins, me smurt)

I know it's not C/Arduino code. It was just to illustrate the logic.

[straight_stewie]

22 hours ago, grimreeper132 said:

How do I split a floating number up into 4 digits, so I can use each one individually so I can display using LEDs. The number I am getting are 25.12 for example which would relate to 2 Red LEDs, 5 Yellow, 1 white and 2 Blue. I can do the part after getting the numbers into the 4 digits split up. The only way I can think of its through a massively long bit of code of ifs going through all possibilities which just seams stupid and that there is a better way to do it. Any Suggestions welcome.

One simpler encoding might be to have each "number" be an array of bytes, each byte representing whichever color. The trick here is, since there are only four colors, this is one 32 bit integer, and so you can do bulk math on the whole value by taking it as an integer, and piecemeal operations by working with the individual bytes.

To get the individual bytes out of the integer you can use bitshifting. For example, if you want the highest byte you can bitshift right 8*3 bits, to get the lowest byte you will first bit shift left 8*3 bits and then bit shift right 8*3 bits. To get a value in between you will bitshift left 8*n and then right 8*3 bits, where n is the non zero based index of the byte you want to get out of the integer.

Another solution is to use bitwise XOR, a bitmask, and a short decision tree to get the bytes out that you want. In both methods I will leave you to figure out how to place the modified bytes back into the integer correctly.

To answer your specific question, the easiest method would be to cast it as a string, and then cast the individual characters of the string to ints, do the math, and then do the reverse.