|
From: Eric F. <ef...@ha...> - 2005-10-03 22:48:58
|
Tim Churches wrote: > Eric Firing wrote: > >>Paul, Tim, >> >>Masked arrays *are* in scipy_core. >> >>import scipy.base.ma as ma > > > OK, thanks. In the absence of documentation, I just looked for an MA > subdirectory, couldn't find one and assumed that it wasn't (yet) supported. > > >>In addition, a quick look indicates that NaN-handling is in good shape. > > > How about other IEEE special values? > > Tim C Tim, It has inf: >>> import scipy.base as nx >>> nx.inf inf >>> nx.isposinf(nx.inf) True >>> nx.isposinf(-nx.inf) False >>> nx.isneginf(-nx.inf) True >>> I don't know about signed floating point zero; I have never used such a beast, and am aware of it only because it is in numarray's ieeespecial module. See http://numeric.scipy.org/new_features.html; it includes: Errors are handled through the IEEE floating point status flags and there is flexibility on a per function / module / builtin level for handling these errors. I haven't looked into how this is done. Eric |