plotting dates over continuous numeric data

So I am using ODE45 to solve a system of equations, and then I plot the solution. The problem I am having is that for each y value, the corresponding x value is a decimal in most cases. Therefore the datetime function is not working, and I cannot seem to edit the xtick values to account for it.

z=0;
Z=0;
i = [ 1 2 3 4 5];
it =[0 28 56 91 119];
j = [3 4 5 6 7];
jt = [0 28 56 91 119];
kt = [8 9 8 7 6];
k = [0 28 56 91 119];
TSPAN = [0 300];
ICC = 500;
for P_1 = [1]
    for P_2 = [2]
        for  P_3 = [3]
                [t,y] = ode45(@(t,y) ODE(t,y,P_1,P_2,P_3,it,i,jt,j,kt,k), TSPAN, ICC);
                z=z+1;
                Z=Z+1;
                OutT{z,:} = [t];
                OutY{Z,:} = [y];

          end
      end
  end
  plot(t,y);

with the ode file

function dydt = ODE(t,y,P_1,P_2,P_3,it,i,jt,j,k,kt)
i = interp1(it, i, t, 'linear', 'extrap');
j = interp1(jt, j, t, 'linear', 'extrap'); 
k = interp1(kt, k, t, 'linear', 'extrap'); 
dydt = P_1*i-P_2*j*y-P_3*k*y;
end

The dates that I want to plot correspond to [0 28 56 91 119], and they are;

   07-Jul-2015
   04-Aug-2015
   01-Sep-2015
   06-Oct-2015
   03-Nov-2015

Any ideas how I can get this to work?

Answer 1

[Star Strider]

Converting your ‘t’ values to date numbers and then to datetime or similar formats is likely going to be an unproductive use of your time, unless they absolutely all must be datetime variables.

I would just do something like this:

Dc = {'07-Jul-2015', '04-Aug-2015', '01-Sep-2015', '06-Oct-2015', '03-Nov-2015'};

figure
plot(t,y);
grid
set(gca, 'XTickLabel',[])
text(jt, -10*ones(size(jt)), Dc, 'Rotation',30, 'FontSize',8, 'HorizontalAlignment','right')

Answer 2

[Peter Perkins]

If you have this

>> x = [0 28 56 91 119]
x =
     0    28    56    91   119

why not just plot your y values against this?

>> d = datetime(2015,7,7) + caldays(x)
d = 
  1×5 datetime array
   07-Jul-2015   04-Aug-2015   01-Sep-2015   06-Oct-2015   03-Nov-2015

It's probably a bit more complicated than that, but probably not much: I guess that x vector is the ticks you want. t is, I guess, the full vector of x values, and they likely include fractional parts? But if t is relative to midnight of 7-Jul, you can plot against d = datetime(2015,7,7) + days(t), because days will accept fractional values (it assumes you mean "24 hour days" whereas caldays allows for the fact that days might also be 23 or 25 hours, and therefore a fraction is not well-defined).

Then set the ticks to the shorter datetime vector.