Array[1,dim] get's changed to Array[dim,1]

[Andreas Heuermann]

I have a model that calls a built-in function point-wise on a vector input with dimension [1,dim], but the output of that function is of dimension [dim,1].

setCommandLineOptions("-d=newInst,optdaedump");
loadString("
model mwe 
  Real vectorA[1,4];
equation
  vectorA = tanh( {{10,20,30,40}} );
end mwe;
"); getErrorString();
simulate(mwe); getErrorString();

When I dump the equation at the first point in the frontend with optdaedump I'll get

Equations (1, 4)
========================================
1/1 (4): vectorA = {{tanh(10.0)}, {tanh(20.0)}, {tanh(30.0)}, {tanh(40.0)}}   [dynamic |0|0|0|0|]

but I would expect to get the same dimension as the input was, so:

vectorA = {{tanh(10.0), tanh(20.0), tanh(30.0), tanh(40.0)}}

This prevents me from solving related ticket #6266.

@mahge930, @perost Any idea where to look at for this?

By the way, if I use the old frontend it gets flattened out is completely wrong.

[Andreas Heuermann]

Using

model mwe 
  Real vectorA[4];
equation
  vectorA = tanh( {10,20,30,40} );
end mwe;

works (when using the newFrontend) as expected.

[Andreas Heuermann]

I guess it is just transposed. When looking at a second test case

model mwe2
  Real vectorA[2,4];
  function testFunc
    input Real x_in;
    output Real x_out;
  algorithm
    x_out := tanh(x_in);
  end testFunc;
equation
  vectorA = testFunc( {{10,20,30,40}, {11,21,31,41}} );
end mwe2;

I get

vectorA = {{mwe2.testFunc(10.0), mwe2.testFunc(11.0)}, {mwe2.testFunc(20.0), mwe2.testFunc(21.0)}, {mwe2.testFunc(30.0), mwe2.testFunc(31.0)}, {mwe2.testFunc(40.0), mwe2.testFunc(41.0)}}

[Mahder Alemseged Gebremedhin]

The input to the function is a matrix (1x4). So vectorization of the function call will give you back a matrix (also 1x4).

I think this is the expected behavior. Right?

[Mahder Alemseged Gebremedhin]

Ahh nvm. You are getting a 4x1 matrix. I see now. Maybe a listReverse missing somewhere.