MATLAB: How to marshal to MATLAB COM Interop

cominterop

Hello,

I'm pretty sure I am seeing a marshaling issue interfacing with MATLAB. I've got a client wiring up just fine in a C# program, and want to call fft, I gather through the Feval call.

Question (or questions) is (are), how to specify and/or marshal the requisite inputs and outputs in such a way that we get the desired response.

There' a follow on call in many of the fft examples to linspace, but I gather that technically this is optional, unless we want to wire up for visualization or otherwise.

I've verified on the "simple" hello world scenario, I call Feval to "version" and obtain the desired response. Now I want to step it up a notch for fft, and so on.

Next from that, various bandpass filters.

I should also mention, we're running against R2011b.

Thank you.

Best Answer

Working through the examples via COM Interop, I am able to make a similar call to the "linspace".

matlab.Feval(@"linspace", 1, out lso, 0d, 1d, nfft / 2 + 1);

And with a bit of C# extension method and lambda persuasion, extract out the desired result.

So at this point I am thinking it's something in how arrays are being passed into MATLAB.

Am I in the ball park if I assume (which I am loathe to do) that it prefers arrays to be 2-dimensional, even if we're talking about a "single dimension" of sampled data? So in other words, instead of double[] (C#), I should be transforming that into double[,]?

Thank you...

Related Solutions

MATLAB: How to convert an MWArray returned by the MATLAB Builder NE component to a native .NET array in Visual Basic

ToArray() is a method that is valid only for the MWNumericArray, MWLogicalArray, and MWCharArray types. An element of an MWArray object must be cast to one of these types first before a ToArray() may be issued.

Use the CTYPE Visual Basic function to cast each element of the MWArray array into an MWNumericArray, MWLogicalArray, or MWCharArray, and then call ToArray() on the result. For example, if 'foo' is the MWArray object that contains the various output arguments, the following statement converts the first output argument into a native .NET Double array by way of the MWNumericArray type:

Dim nativeArray(,) As Double = CType(foo(1), MWNumericArray).ToArray(MWArrayComponent.Real)

Refer to the MATLAB Builder NE MWArray Class Library Reference for more details:

https://www.mathworks.com/help/compiler_sdk/cxx/mwarray.html

MATLAB: Is there an example on how to use FEVAL from C# code

Below is an example showing the use of FEVAL from C# code:

add these two lines at the beginning

using System;
using System.Reflection;
      // Create or get an instance of MATLAB
      Type matlabType = Type.GetTypeFromProgID("Matlab.Application");
      object matlab = Activator.CreateInstance(matlabType);
      // Object[] parameters = new Object[] { functionName, nargout, Result, arg1, arg2, ... argN };
      Object[] parameters = new Object[] { "strcat", 1, null, "Hello", " World!" };
      // Make 3rd parameter (the return value) be passed by reference
      ParameterModifier mods = new ParameterModifier(parameters.Length);
      mods[2] = true;
      try
      {
          // Call Feval to evaluate the result of "strcat" function 
          matlabType.InvokeMember("Feval",
                                  BindingFlags.InvokeMethod,
                                  null,
                                  matlab,
                                  parameters,
                                  new ParameterModifier[] { mods },
                                  null,
                                  null);
          // Get the return value
          object Result = ((object[])parameters[2])[0];
          // Print the return value
          Console.WriteLine(Result);
      }
      catch (TargetInvocationException e)
      {
          Console.WriteLine(e.Message);
      }

Best Answer

Related Solutions

MATLAB: How to convert an MWArray returned by the MATLAB Builder NE component to a native .NET array in Visual Basic

MATLAB: Is there an example on how to use FEVAL from C# code

Related Question