|
From: Francesc A. <fa...@ca...> - 2005-11-10 13:04:01
|
A Dijous 10 Novembre 2005 13:34, Niklas Volbers va escriure: > > I think you should take a look at PyTables for more elegant solutions. > > I wasn't sure if it will stay compatible to scipy_core, because right now > it uses numarray. But from what you said I am sure that eventually > pytables might also switch to scipy_core. Just to make things clear, PyTables uses numarray at its core, yes, but it does support Numeric also by doing conversions in the fastest way available (in fact, many people uses PyTables just for keeping Numeric arrays, with good results). Still, a copy in memory is needed for numarray-->Numeric conversions (not for the Numeric-->numarray way), but that will change very soon, after the array protocol implementation in numarray (and Numeric) stabilizes. The next step will be to offer support for scipy.core by using the array protocol as well. And finally, we plan (probably with the introduction of PyTables 2) to switch from numarray to scipy.core when the later would be stable enough, but always with support of the triad of numarray, Numeric and scipy.core, at least for a few years. Nevertheless, you must be aware that PyTables does require the HDF5 libraries [1], so, if what you want is to get rid of the maximum of dependencies, PyTables might be effectively overkill for you. [1] http://hdf.ncsa.uiuc.edu/HDF5/ Cheers, =2D-=20 >0,0< Francesc Altet =A0 =A0 http://www.carabos.com/ V V C=E1rabos Coop. V. =A0=A0Enjoy Data "-" |