SerializedField float in inspector automatically rounds (Solved)

Hi everyone,

I'm currently working on a project that needs to handle GPS latitude and longitude values.
I have an object with a script attached that should hold the latitude/longitude values assigned to this object.

Code (csharp):

1. [SerializeField]
2. [Tooltip("Set the longitude for the center of the object")]
3. private double _poiCenterLongitude = 0.0f;

Now if I enter a value like this: 48.37201 in the inspector it get's automatically rounded to 48.372. If I then output the value via

Code (csharp):

1. Debug.Log(_poiCenterLongitude);

I get the rounded value as well. Is there any way to stop the inspector from messing with these values? Except changing the type to double? I think float should be enough (shouldn't take values larger than this, also Input.location.lastData.longitude is a float as well).

Thanks in advance. I tried searching for a solution but didn't find anything useful.

-lord_of_shred-

 

The problem here is not unity but the way floating point numbers are stored.
I guess the number you mentioned is not representable by a float. I didn't do math to proof this, but did some tests with that number myself.
So, I guess you will need doubles if you require such a precision.

 

Unfortunately math doesn't care what anybody thinks. With a standard 32-bit

float

, you are actually only guaranteed about six decimal digits. Reviewing the details on wikipedia,

https://en.wikipedia.org/wiki/Single-precision_floating-point_format

I quote, "If a decimal string with at most 6 significant digits is converted to IEEE 754 single-precision representation, and then converted back to a decimal string with the same number of digits, the final result should match the original string."

Your example of

48.37201

is actually 7 significant digits.

Why not just store the coordinates in a string?

 

IDK.

Code (CSharp):

1. using UnityEngine;
2.  
3. public class KeyCodeTest : MonoBehaviour
4. {
5.     [SerializeField]
6.     private double test;
7.  
8.     private void Start()
9.     {
10.         Debug.Log(test);
11.     }
12. }
13.  

And after hitting play.


2019.3.14f1
Did I misunderstand something? (Don't mind the class name )

 

We're talking about

float

types, which are 4-byte. In the original post, @lord_of_shred indicated he was not interested in

double

, that's all.

 

Ah, I'm having code-fixation. I haven't seen the tailing question, only the code. Sorry, I'll see myself out.

 

S'all good my friend... as a proper Ninja, you're welcome to hang out and nobody will notice you in the shadows anyway, right??

 

I think I'd just use an

int

to represent a fixed point longitude/latitude value with 6 decimals accuracy, so
48.37201 would be stored as 48372010. You could even go up to 7 decimals if you wanted.
There is probably a way to use a custom editor to still show the decimal point in the inspector, but I'd probably just make it plain in the variable names and call them e.g.

latitudeMicros

.

 

Thanks for all your replies, I'll set this to (Solved).
I went with doubles to avoid further troubles with it.

You're right, math doesn't care ;-). I had thought about the IEEE 754 "restrictions" but dismissed it, since Unity seems to work with float for the GPS readout.