How to convert String to Double ?

I am trying to convert a string array into the double array.

var = 1×11 string    
"ux"         "uy"         "t"          "u"          "theta"      "x"          "y"          "r"          "R"          "H"          "th"         

This is the string array i want to convert in double array

Required output:

var = 1×11 double 
ux uy t u theta x y r R H th

I tried the following code :

vard=str2double(var)
var=vard

And this is its output

var = 1×10    
   NaN   NaN   NaN   NaN   NaN   NaN   NaN   NaN   NaN   NaN

Thanks for your time and consideration.

Thanks in advance for helping me to solve this issue.

[Adam Danz]

Are the strings variables names? In other words, are you trying to convert "ux" into the number stored by variable ux?

[Rushil Shah]

Yes I am using them as variables.

Here's the entire code.

list=["ux","uy","t","u","theta","x","y","r","R","H","th"]
% indx is index of the variables chosen by the user
% answer is the array which contains value of the variables chosen by the user
var=list(indx);
val=answer;
for n=1:numel(var) % this is the code to assign value from answer to the variables created from var string array
vn=val(n);
assignin('base',char(var(n)),vn{:})
end
class(var)

Here every variable has 'char' datatype which can't be used for calculations.

That's why I need them to convert it into double format.

Answer 1

[Walter Roberson]

You cannot. Names such as ux can never be data class double. The closest you could get would be categorical(var)

If what you mean is that you have a string array of names, and for each name there is variable with that name in the workspace, and you wish to replace each name with the content of the variable with that name, then if you have the Symbolic Toolbox, use subs()

[Rushil Shah]

Thank you so much for your reply !

Yes I am using them as variables.

Here's the entire code.

list=["ux","uy","t","u","theta","x","y","r","R","H","th"]
% indx is index of the variables chosen by the user
% answer is the array which contains value of the variables chosen by the user
var=list(indx);
val=answer;
for n=1:numel(var) % this is the code to assign value from answer to the variables created from var string array
vn=val(n);
assignin('base',char(var(n)),vn{:})
end
class(var)

Here every variable has 'char' datatype which can't be used for calculations.

That's why I need them to convert it into double format.

[Rushil Shah]

I am making a code to solve n number of equation by itself.

So for that i'm asking user to choose 2 or more variables and enter their values.

Here I'm solving 11 equations of Projectile Motion

I have 11 variables and 11 equations.

I need user to choose any 2 variables and enter its value and get 11-2=9 answers by solving those 11 equations.

list=["ux","uy","t","u","theta","x","y","r","R","H","th"] % variables
[indx,tf] = listdlg("ListString",list,"SelectionMode","multiple")
indx % shows which index number is chosen by that we will know which variable is chosen
list1=["Enter a value of ux in m/s"; "Enter a value of uy in m/s"; "Enter a value of t in s"; "Enter a value of u in m/s"; "Enter a value of \theta (in degrees)"; "Enter a value of x in m"; "Enter a value of y in m"; "Enter a value of r in m"; "Enter a value of R in m"; "Enter a value of H in m"; "Enter a value of th in s"];
list1(:,1)=''''+list1(:,1)+''','
prompt=list1(indx);
dlgtitle = 'Input Value';
definput = {'0','0','0','0','0','0','0','0','0','0','0'};
opts.Interpreter = 'tex';
answer= inputdlg(prompt,dlgtitle,[1 40],definput,opts);
% answer array has the values of the variables

syms ux uy t u theta x y r R H th
g=-9.81

% creating chosen variable from list string array and assigniong corresponding value from answer array
var=list(indx);
val=answer;
for n=1:numel(var)
vn=val(n);
assignin('base',char(var(n)),vn{:})
end
var
class(var) % output is STRING
str2double(var) % output is NAN

% for i=1:numel(var)
%     var(i)=str2double(var(i))
% end
% Above code didn't work and shown the same result NAN

% Below code worked once but then it also created some issue
% I used switch case because I just Wanted to convert the chosen variables from string to double 
% For loop didn't work in so wasn't able to perform the same action on all variables
class(indx)
in=cast(indx,'char')
vardouble=str2double(var)
switch in
    case 1
        ux=str2double(ux)
    case 2
        uy=str2double(uy)
    case 3
        t=str2double(t)
    case 4
        u=str2double(u)
    case 5
        theta=str2double(theta)    
    case 6
        x=str2double(x)    
    case 7
        y=str2double(y)
    case 8
        r=str2double(r)
    case 9
        R=str2double(R)
    case 10
        H=str2double(H)
    case 11
        th=str2double(th)
end

% this didn't work manually
% in general we need to convert every chosen variable's datatype (class) from string to double
R=str2double(R) % output is NAN

% u=10                                     for Testing purpose
% theta=15                                 for Testing purpose
% R=5.0968399592252803261977573904179      for Testing purpose
% H=0.34142353775627248824164894358546     for Testing purpose

% eqns=[theta==acosd(ux/u),ux == u*cosd(theta),uy == u*sind(theta)-g*t,u==sqrt((ux.^2+uy.^2)),t==(2*u*sind(theta))/g,x==u*t*cosd(theta),y==u*t*sind(theta)-(g*t.^2)/2,r==sqrt((x.^2+y.^2)),th==(u*sind(theta))/g,H==(u.^2*sind(theta)^2)/(2*g),R==(u.^2*sind(2*theta))/g]

eqns1=[H==(u.^2*sind(theta)^2)/(2*g),R==(u.^2*sind(2*theta))/g,t==(2*u*sind(theta))/g,th==(u*sind(theta))/g]
% eqns1 is for Input R and H only
% in general we need all equations as shown in eqns

AB = vpasolve(eqns1,[u theta t th])
% [u theta t th] is for Input R and H only
% in general we need every other variables which isn't chosen by user

% in general we need every other variables which isn't chosen by user for output
AB.t
AB.th
AB.u
AB.theta

[Walter Roberson]

list = ["ux","uy","t","u","theta","x","y","r","R","H","th"] % variables

var1idx = menu('Pick a variable', list);
var1 = list(var1idx);
var1val = input("enter a value for " + var1 + " ? ");

list2 = list;
list2(var1idx) = [];
var2idx = menu('Pick a second variable', list2);
var2 = list2(var2idx);
var2val = input("enter a value for " + var2 + " ? ");

symlist = sym(list);
vallist = subs(symlist, {var1, var2}, {var1val, var2val});

new_equations = subs(equations, symlist, vallist);

Now you can solve the 11 new_equations in 9 variables.

But you probably will not get back 9 solutions each, or 9 solutions. You probably will not be able to get back more than 7 solutions.

[Rushil Shah]

Thanks for your positive respose.

I tried your code but it is showing error and not giving any solutions back. I love the idea of using 'menu' function. 'vpasolve' is working better and faster than this method. We need to solve 11 equations so subsituting the vaues isn't working.

CODE

syms ux uy t u theta x y r R H th
g=-9.81;
eqns=[theta==acosd(ux/u),ux == u*cosd(theta),uy == u*sind(theta)-g*t,u==sqrt((ux.^2+uy.^2)),t==(2*u*sind(theta))/g,x==u*t*cosd(theta),y==u*t*sind(theta)-(g*t.^2)/2,r==sqrt((x.^2+y.^2)),th==(u*sind(theta))/g,H==(u.^2*sind(theta)^2)/(2*g),R==(u.^2*sind(2*theta))/g]

list = ["ux","uy","t","u","theta","x","y","r","R","H","th"] % variables

var1idx = menu('Pick a variable', list);
var1 = list(var1idx);
var1val = input("enter a value for " + var1 + " ? ");

list2 = list;
list2(var1idx) = [];
var2idx = menu('Pick a second variable', list2);
var2 = list2(var2idx);
var2val = input("enter a value for " + var2 + " ? ");

symlist = sym(list);
vallist = subs(symlist, {var1, var2}, {var1val, var2val});
sol=solve(eqns,[ux uy t th r R H u theta x y])
% equations=[H==(u.^2*sind(theta)^2)/(2*g),R==(u.^2*sind(2*theta))/g,t==(2*u*sind(theta))/g,th==(u*sind(theta))/g]

% AB = vpasolve(eqns1,[u theta t th])
new_equations = subs(sol, symlist, vallist);
new_equations

ERROR SHOWN

Warning: 
Unable to solve symbolically. 
Returning a numeric solution using vpasolve.

Error using subs
Expected input number 1, S, to be one of these types:

sym

Instead its type was struct.

Error in sym/subs (line 60)
validateattributes(F, {'sym'}, {}, 'subs', 'S', 1);

Please help to change the datatype of those variables.

[Walter Roberson]

Returning a numeric solution using vpasolve.

You solve()'d all 11 original equations for all 11 variables. MATLAB was not able to find all the exact solutions, or even one exact solution, so it returned one numeric solution. That numeric solution will not have any variables in it. Numeric solutions returned by solve() or vpasolve() never have remaining variables in them.

The only time you can get a solution with remaining variables is if you solve() for fewer variables than you have equations and solve() is able to find an analytic form for the variables you solve for in terms of the remaining variables.

new_equations = subs(sol, symlist, vallist);

sol is your struct that has a single numeric solution. It has no variables in it to have their variables substituted. But if you insist on trying, then

new_sol = structfun(@(S) subs(S, symlist, vallist), sol, 'uniform', 0);

You will see that the result is exactly the same set of numeric values as the original struct.

We need to solve 11 equations so subsituting the vaues isn't working.

You have 11 equations in 11 variables. If there are solutions at all, then it will only be when the values of the variables have very precise relationships to each other. But you are having the user enter 2 arbitrary numeric values for variables, using whatever values that they want, and those arbitrary numeric values that they enter will almost certainly not have the necessary precise relationships to each other. There will be no solution nearly all of the time.

If you want to be able to have the user provide exact values for two of the variables, then you need to remove two of the equations, each involving at least one of the two variables.

Because you have trig in your equations, there is something meaningful you can do with user-provided values: you can use the user-provided values as starting points for vpasolve. Not as exact values, but as starting points to help find a close-by period of the values. For example,

symlist = sym(list);
vallist = subs(symlist, {var1, var2}, {var1val, var2val});
vallist(ismember(vallist, symvar(vallist))) = nan;

sol = vpasolve(eqns, symlist, vallist(:));

Due to the way that vpasolve() finds values, the solution that it finds will not necessarily be the one "closest" to the initial values that the user provides.

And to emphasize, the solutions found will almost certainly not have the corresponding variable have exactly the value provided by the user.

Consider for example vpasolve(x^2+sin(x)==4, x) then there are multiple solutions, and if you have the user provide an initial value, vpasolve(x^2==4, x, INITIALVALUE) then if the INITIALVALUE is positive then typically a solution near +2 would be found, and if the INITIALVALUE is negative then typically a solution near -2 would be found: the user is not providing the exact solution, just a clue to vpasolve() of where to look, and vpasolve() might or might not be able to take advantage of that clue.