MATLAB digits precision

0
votes

I have MATLAB code:

clear;
clc;
syms x;
f=log(x)*sin(x^2);
a=vpa(subs(f,x,2),100)
fprintf('a=%.100f\n',a);
doublea=double(a);
fprintf('a=%.100f\n',doublea);

and the result is

a =
 
-0.5245755158634217064842071630254785076113576311088295152384038229263081153172372089356742060202648499
 
a=-0.5245755158634216600000000000000000000000000000000000000000000000000000000000000000000000000000000000

doublea =

   -0.5246

a=-0.5245755158634216600000000000000000000000000000000000000000000000000000000000000000000000000000000000
>>

Why if I use fprintf the decimal precision is only up to 16 digits, even though I use 100 digits precision? Also why if I convert a to double then a is only up to 16 digits?

Can it cause an error in the calculation if I want use more than 16 digits precision? How to fix it?

1
matlabprecision
this should explain :stackoverflow.com/a/4227530/3878321 - Hadi
fprintf converts the vpa number to double to print it. A double has about 16 decimal digits of precision, it cannot store more digits. If you need more, you need to use vpa. Use disp to display its value (or use char to convert to string, and display that). - Cris Luengo

1 Answers

0
votes

Have you tried using num2str or vpa coupled with disp?

a1 = vpa(a,100);
disp(a1);

Or

disp(['a = ' num2str(a,100)])

Edit: Hadi's link in the comments has a great explanation.