Unexpected end of line - Mata not going through loops

[Tor Madsen]

Hi, I'm currently doing an introductory econometrics class.

In this assignment we have to do 1000 simulations on three different sample sizes.
This is done via two for-loops (Ts = column vector containing the three sample sizes):

for(j=1; j<=length(Ts);j++) {
    // select the current sample size
    T = Ts[1,j]
    
    //loop for the replications
    for(i=1; i<=M; i++) {
        
        // random number generation
        y = AR_gen(1,T)
        x = AR_gen(1,T)
        
        //OLS regression
        X = J(T,1,1), x
        aux_ols = luinv(cross(X,X)*cross(X,y)
        beta[i,j] = aux_ols[2,1]
        
        // tstat
        err = y-X*aux_ols
        RSS = cross(err,err)
        sigma = luinv(cross(X,X))*RSS/(T-2)
        tstat[i,j] = beta[i,j]/sqrt(sigma[2,2])
        
        // R2
        TSS = cross(y-mean(y)*J(T,1,1),y-mean(y)*J(T,1,1))
        R2[i,j] = 1- RSS/TSS
        
        // DW
        DW[i,j] = cross(err[2::T,1]-err[1::T-1,1],err[2::T,1]-err[1::T-1,1])/RSS
        
    }
}

where the AR_gen function is made previous in the do.file:

function AR_gen(phi, T) {

c = J(T+1,1,0)
eps = rnormal(T,1,0,1)

c[1,1] = 0 // 0, 0.S/(1-phi)

for (i = 1; i<=T; i++) {
    c[i+1,1] = phi*c[i,1]+eps[i,1]
}

v = c[2::T+1,1]

return(v)

}

The do.file starts with the 'mata' command and ends with 'end'.
When i run the do.file it instantly reports back 'unexpected end of line', but it seems that everything before the loop has been processed.
It saved all the zero-matrices i made to store the betas, tstats, DW and R2 in.

Im quite a novice at both Mata and econometrics and I just can't figure out where i made a mistake or why mata doesn't try to iterate the actions in the loops

Any help appreciated

[Mike Lacy]

One mistake I can spot is that this line:

aux_ols = luinv(cross(X,X)*cross(X,y)

has unbalanced parentheses. That error, for me, is a common reason for messages like "unexpected end of line."

A simple debugging trick for Mata programs is to print out something at several points in the code, thereby letting you narrow down exactly which line is creating the problem. I have in mind something like this:

Several lines of mata code ....
"1"
More mata code ...
" 2"
More code yet ...
"3"

Running the code will let you narrow down the error to (say) between 1 and 2. At that point, you can add in another "1.5" or whatever to further narrow down the location of the error, and so forth.

[Tor Madsen]

Hi, I'm a bit late on this one but thanks for your help, the entire problem was exactly that one missing paranthese.

The debugging trick sounds like a good idea, but it doesn't seem to do anything when the problem lies within a loop.
In this case mata didn't even begin to execute the loop but everything before it worked fine.

Anyway thanks