Data_oM: Multi-dimensional arrays

[tg359]

Description:

We have to do a lot of maths with lists, arrays and matrices around results processing, running parallelisable calculations and transforming datasets from flat files into n-dimensional arrays to run some funky numerical analyses. It would be nice to have some kind of numeric array object that these kinds of operations can be applied to (think numpy.ndarray and its associated functionality).

What are others thoughts on this? It would be great to get a load of the methods and workflows already built in other languages in here to make them more accessible!

We might be able to use some of the Numpy.NET stuff if possible, as the majority of operations are already in that (though .net version might prevent us from doing this).

The core functions that are needed for what we currently do would be:

• element-wise operations between arrays of the same shape (+, -, *, /)
• array reshaping (n*m > m*n, or l*m*n > m*l*n)
• array-wise operations (applying +, -, *, / to each of the elements in an array)
• axis-wise operations (sum, min, mean, quantile, max)
• higher level operations - nD-interpolation