[RomanovaDI]

I need to convert volTensorField variable tf of OpenFOAM to c++ array of double. I've got the error:

cannot convert ‘Foam::volTensorField’ {aka ‘Foam::GeometricField<Foam::Tensor<double>, Foam::fvPatchField, Foam::volMesh>’} to ‘double*’

If I use tf.cdata() the error is:

cannot convert ‘const Foam::Tensor<double>*’ to ‘double*’

[Tobermory]

Well that is understandable. A volTensorField is a completely different object to an array of doubles ...

Not convinced? Well, volTensorField is a typedef in volFieldsFwd.H (https://cpp.openfoam.org/v8/volField....html#l00062):

typedef GeometricField<tensor, fvPatchField, volMesh> volTensorField;

i.e. is a GeometricField. Now take a look in GeometricField.H (https://cpp.openfoam.org/v8/Geometri...8H_source.html) and you'll start to see the difference.

[olesen]

Originally Posted by RomanovaDI

I need to convert volTensorField variable tf of OpenFOAM to c++ array of double. I've got the error:

cannot convert ‘Foam::volTensorField’ {aka ‘Foam::GeometricField<Foam::Tensor<double>, Foam::fvPatchField, Foam::volMesh>’} to ‘double*’

If I use tf.cdata() the error is:

cannot convert ‘const Foam::Tensor<double>*’ to ‘double*’

If you don't mind some ugliness and ONLY do this correctly, you can write something like the following:

volTensorField fld = ...;


reinterpret_cast<const scalar*>(fld.primitiveField().cdata());


// OR 
reinterpret_cast<const pTraits<tensor>::cmptType*>(fld.primitiveField().cdata());

but you must be really, really careful and know exactly what you are doing.
NOTE: always use scalar instead of double for the underlying component type since it is also possible to compile with single-precision or mixed-precision and then the tensor will comprise float not double.


NOTE: the above will only get you read access to the internal volume field, which is likely what you want anyhow. No really easy to do the same thing for each of the boundary fields as well without lots of code.