Following the previous post the Scipy statistics functions have been updated to include all the distributions, including the discrete distributions. The input to the distribution functions has also been updated to allow arrays to be transferred either as Excel ranges or pyxll cache objects. The updated code and spreadsheets can be downloaded from:

py_SciPy.zip

The screenshot below shows the py_multivariate_norm function with the x array input as a cache object:

Changing x to the range G32:G36 returns exactly the same results:

There have also been changes to the py_linspace function, in the pyNumpy.py file, to allow the results array to be returned either as a range or a pyxll cache.

If the optional Rtn_Cache input is set to False (default) the results are returned as a range:

If Rtn_Cache is set to True the array is returned as a pyxll cache, as for py_linspace_o, which is still available:

The updated pyNumpy.py file comes in two versions, pyNumpy-jit.py and pyNumpy-noJIT.py, with the jit version including calls to the Numba just-in-time compiler. The chosen version should be copied to pyNumpy.py.