Is it impossible to multiply matrices/vectors by scalar values in Stan?

Question

Consider the below Stan script, which I'm using via RStan:

data {
  int K; //outcome classes
  int N; //rows
  int D; //input dimensions
  int y[N];
  matrix[N, D] X;
  real days[N]; 
}
parameters {
  matrix[D, K] C;
  matrix[D, K] B;
}
model {
  matrix[N, K] pred = X*C + days*X*B; //If I remove days, it works fine. 

  to_vector(pred) ~ normal(0, 5);

  for (n in 1:N)
    y[n] ~ categorical_logit(pred[n]');
}

An undesirable solution, is just duplicating the days column in R such that it IS a matrix and can be used in matrix multiplication. But is that really necessary? Should be easy to "scale" a matrix by a scalar value.

The error is:

Edit: I've also cast days as a matrix matrix[N,1] days[N]. While this does pass the "scrub" it gets rejected at compilation due to dimension mismatch.

Edit2: I've adjusted the code, which does now execute. But I'm confused why this nested for loop is even necessary. There should be an easy way to multiply all vector elements by the same scalar value.

data {
  int K; //outcome classes, 3
  int N; //num rows
  int D; //input dimensions, 5
  int Y[N];
  matrix[N,D] X;
  int days[N]; 
}
parameters {
  matrix[D, K] C; //[5,3] 
  matrix[D, K] B; //[5,3]
}
model {
  for (n in 1:N){
    vector[K] pred;
    vector[D] ipt;
    matrix[K,K] day_diag;
    for (i in 1:K){
      for (j in 1:K){
        if (i == j)
          day_diag[i,j] = days[n];
        else
          day_diag[i,j] = 0; 
      }
    }
    
    ipt = X[n]'; // now row_vector [1xD]
    // [D,K] x [1,D] + [D,K] x [1,D]  
    pred = C * ipt   +  B * (day_diag * ipt); 
    Y[n]~categorical_logit(pred);   
  }
}

Answer 1

To scale each row of the matrix by the corresponding value of a vector, you can use diag_pre_multiply(), so it would be diag_pre_multiply(days, B). days will need to be read in as vector[N] days rather than real.

As an aside, Stan has a tonne of matrix operations defined, see: https://mc-stan.org/docs/2_25/functions-reference/matrix-operations.html