I am having lots of strings that represent rotation values. These strings have 0 to 5 decimal places.
When converting these strings to floats to apply them as rotation values to GameObjects, the float value is always cut short to 4 decimal places. With more than 500 GameObjects, this will result in a huge imprecision over time.
How can I preserve all decimal places?
This is the code I used [modified for better understanding]:
myString = "344.93873";
segment.transform.rotation = Quaternion.Euler((float)double.Parse(myString), 0f, 0f);
What I get is 344.9387 without the 3 at the end. The loss happens here: (float)double.
However, when manually rotating objects in Unity3D, I can see rotation values with more than 4 decimal places in the Inspector.
Per default float.ToString
The ToString() method formats a Single value in the default ("G", or general) format of the current culture.
And further per default for float (=Single)
Single
Smallest round-trippable number of digits to represent the number (in .NET Framework, G7 is the default)
Now your input number
344.93873
has 8 digits so per default it is clamped to
344.9387
which has only 7 digits.
A float can have ~6-9 digits so in order to be sure to always see the full value you should use
print(myFloat.ToString("G9"));
Besides that: Yes indeed have in mind that either way floating point numbers have a limited precision.
It is very much possible that the exact value of
344.93873
is not representable using a float!
See .Net fiddle
double.Parse() => 344.93873000000002093656803481280803680419921875
float.Parse => 344.938720703125
Tbh honest though in your use case of angles. There is no way whatsoever, not even in nature anyone will notice a difference of a 0.00001° angle!
Have in mind that after all Unity is a realtime 3D "Game"-Engine which on purpose drops precision in favor for a faster runtime.