MaxCoderz

This is quite annoying, and google isn't cooperating.

Suppose we have an array of data, called "data", and an array of incides called "indices" (very original..)
All indices [0 .. n-1] are present, and only once.

The array "result" is such that for every n, result[n] equals data[indices[n]].

Can the result be generated without copying the items of "data" into a new array? If so, how?
In other words, can "data" be modified in-place to have the property of "result"? (of course in that case the "data" used at the right hand side is taken from the old array, but that's just a technicality)

I really feel like it ought to be possible..

I can't think of any way that wouldn't require some additional storage where the amount of storage relates to the number of elements, sorry.

What language? Methinks you failed to specify...

Doesn't matter, all I need it an algorithm
But, please don't use that as an excuse to use something exotic

edit: ok, Universe, you have won. Until I find a better way, I will just copy all data to a secondary array.

King Harold wrote:Doesn't matter, all I need it an algorithm
But, please don't use that as an excuse to use something exotic

edit: ok, Universe, you have won. Until I find a better way, I will just copy all data to a secondary array.

http://www.jjj.de/fxt/

Have a look at the permutation functions of the lib, you might find something. Or just google for in-place permutation algorithms...

Cool! thanks