Hi

I’m having some problems with a very simple ode which has a single dimension.

(It’s analytically solvable which I know is preferred, it’s the first piece in a model which is not).

In the examples for using the ODE solver in stan the state always has more than 1 dimension,

e.g. real y[N, 2].

If I want to have y have a single dimension: real y[N] I get a dimension mismatch.

Is this deliberate? Do I need to always define the state and initial conditions as

real y[N, 1]? Or should “real y[N]” work?

Hi. If you’re simulating/fitting an ODE with one state then you have to declare `y`

as you stated in order for the Stan program to parse. For the sake of clarity (in the context of simulation),

```
...
generated quantities {
real y[N,1];
y = integrate_ode_rk45(ode_func, yini t0, time, theta, x_r, x_i);
}
```

If you need `y`

to be of dimension `y[N]`

so that it’s compatible with some other part of the model then you can use the `to_array_1d`

or `to_vector`

function (see the **Mixed Operations** section of the Stan Reference Manual which starts on p.474).