I have real-valued one-dimensional data that I need to approximate using a linear combination of 3 functions. Usually, I would use Fit for that purpose, which uses the least squares method. But, in each data point, in addition to the function value, I have the function derivative value. I need the fitting algorithm to try to match those as well (squares of value differences and square of derivative differences contribute to the total sum of squares being minimized with equal weights).

How can I do that? Can I do use Fit or FindFit for that purpose?

Because my data real-valued, i thought I can represent function and derivative values as real and complex components of complex data points. But, apparently, FindFit does not like complex data.

Fit appears to work with my representation via complex numbers. At least, there are no error messages. I will check if it returns reasonable results.

Looks like the Compiler is finally generating useful stuff: reference.wolfram.com/language/ref/datastructure/Deque.html

Nice to have some of these core data types to mess with

@b3m2a1 How is it related to the compiler?

@C.E. the types were introduced as part of the project and are implemented at the compiler level and can be natively used there

Which means for complicated graph traversal algorithms or things of that nature where an iterative approach is necessary one can use the right data structures for the job and then compile the code for performance

@b3m2a1 I see, that's awesome!