Array derivative

[Oficmajster]

How can I differentiate an array? When I connect the same array to the integral block it automatically understands it as a z(x,y) function. I cannot do the same with the derivative.

What should I do?

Thanks

[Aojha]

In Numerical analysis differentitation means difference between two values devided by the difference in there position.

So try to take difference of two consecutation array values and devide it by index.

[Oficmajster]

And the result will be exactly the same as if I differentiated my z(x,y) function by x?

And by which index I should divide it? If I take value A, with index 10, and value B with index 11, should I just subtract (A-B) and dide it by... 10,5?

[johnsold]

What functions or VIs are you using? The only ones I see are for 1D arrays. Integrating or differentiating a 2D array requires specifying a path or direction through the array. Even in analytic mathematics these operations are usually done one dimension at a time.

Lynn

[Oficmajster]

I need partial derivatives. My array is my Z(x,y) function and I need both dZ/dx and dZ/dy. I'm trying to use derivative x(t).vi and some solutions I found on the Web. 

By specified path you mean that I have to add all raws of my array to make 1D array, and then differentiate, and then again the same with columns?

[altenbach]

There are two basic integration functions (plus a few more).

Integral (x(t): takes a 1D array and outputs a 1D array corresponding to the running integral of the function.

This operation can be approximately reversed using the derivative x(t) function.

BOth only work for 1D arrays.

Numeric integration: Takes a 1D, 2D, or 3D array and returns a single scalar number, corresponding to the integral.

This operation cannot be reversed, because there are infinite possibilities to go from a single scalar to arrays.

[altenbach]

---

@Oficmajster wrote:

I need partial derivatives. My array is my Z(x,y) function and I need both dZ/dx and dZ/dy. I'm trying to use derivative x(t).vi and some solutions I found on the Web. 

---

dZ/dx depends also on y and DZ/dy also depends on x. How do you want the output to look like?

What did you find "on the web"? Do you have a link? (this seems easy enough. Have you tried?)

[Oficmajster]

I know the math principles. Look: say we have a function z(x,y) and we get 2 partial derivatives. Tells the story. My problem is that don't have the function as a formula, I have an 2D array. So I need either to differentiate (somehow) the array, or change it (somehow) into the formula and then proceed. I know about integrals, for I used them in my program previously, now I need partial derivatives.

Thanks for your responses, by the way 🙂

[johnsold]

For example at y = 6 you get a 1D array of z(x). Differentiate that. Repeat for y = 7... Then for each x calculate dz/dy.  With for loops and autoindexing this can be quite straighforward.

Lynn

[altenbach]

As a first approximation, you could take the difference of adjacent values. Your output array would be a complex array where, for each x,y position, the real part is dz/dx and the imaginary part is dz/dy.

.