Hi,

I have a function where the inputs are double* (to the first element of an array, which is coming from Excel VBA).

I have lots of calculations in C++ using Eigen library which are based on floats. I would like to copy these input doubles into my Eigen::VectorXf (which are floats).

What is the best way to do this ? Do I have to loop through the array one-by-one and put them into my Eigen float vector? or is there another way to do this ? a command ?


Please note that I don't know whether the double* inputs from VBA are aligned or not, but I need my VectorXf to be aligned in memory.

Also I would like to have my VectorXf to be const. Do I have to cast to const afterwards ? and how ?

ps: my array sizes are not that big, most of them of size 4, or 16, and some of size 440 doubles.

Thanks in advance.

mehrdadh wrote:I have a function where the inputs are double* (to the first element of an array, which is coming from Excel VBA).

I have lots of calculations in C++ using Eigen library which are based on floats. I would like to copy these input doubles into my Eigen::VectorXf (which are floats).

What is the best way to do this ? Do I have to loop through the array one-by-one and put them into my Eigen float vector? or is there another way to do this ? a command ?

You can use Map<VectorXd> to represent the array of doubles and then use the member function cast() to transform them into floats. The resulting code would look something like this:

Code: Select all

const VectorXf v = Map<VectorXd>(buf, n).cast<float>();


Here, buf is the array you get from Excel VBA and n is the number of elements in the array. By default, Map does not assume the data to be aligned.

Oh thanks so much.

This is awesome

So in this code, my VectorXf, which is called v, is going to be aligned and const.

I think the more correct way is this (having const in Mapping):

Code: Select all

const VectorXf v = Map<const VectorXd>(buf, n).cast<float>();


Perfect! I am going to try it now.