All possible permutations (with repetition) of a 1D array given a new array size

[Mancho00]

---

@altenbach wrote:

---

---

Define "better". 🙂 We are dealing with small numbers, so that loop completes in nanoseconds, basically for free. Power of x requires floating point, thus requiring two coercions (first at the function input and later wired to N) and math on twice as many bytes. Seems ugly. 🙂

Of course you also need to ensure that the loop output is correct for strange inputs. Mine is correct for 0 to the power of 0 equal 1, for example. Of course it will not work for negative powers or anything that uses or results in fractions, etc.

---

Better, as in performance while scaling to larger data sets, as I assume was (part of) the purpose of this exercise.  The coercion ugliness I saw right away, but wasn't sure of the cost at a "large" scale, mostly because I have no idea what is going on inside that function.  The inputs would need to be constrained regardless.

Not that I have much room for argument since my recursive solution would not exactly be "high performance" code at a large scale.  In my defense, I knew there was an easy way to do it, if not the exact solution.  I just chose recursion for fun.

(That's my excuse and I'm sticking to it.):smileyhappy:

[altenbach]

Even on 64bit LabVIEW, FOR loops are limited to the range of I32, same for array sizes. You get there very quickly with unreasonable inputs. 🙂 You bang into the same limit with the power function, so you need to validate the inputs anyway. 

[Mancho00]

It was really a curiosity on my part.  I know you are very good at finding the simplest solution, so it confused me a bit that you would build a loop to do the calculation when there is an existing function.