Is there a way to avoid this inaccuracy when converting an string to a double and then to an unsigned int? I have tried using str2num() but get the same result.Thanks!

uint64(str2double('132697151393876469'))

%ans =

% uint64

% 132697151393876464

% note that the ans is the same for inputs 1326971513938764XX, where XX is in the range (57-71)

Answer 1

[Walter Roberson]

format long g
A = sscanf('132697151393876469', '%lu')

A = uint64
   132697151393876469

class(A)

ans = 'uint64'

Answer 2

[the cyclist]

The reason for this is that MATLAB uses double-precision arithmetic by default, which will typically give approximately 15 decimal digits of precision.

One possibility, if you have the Symbolic Math Toolbox, is to use variable-precision arithmetic.