Hi, On Mon, Apr 06, 2009 at 03:12:38PM +0200, Valentin Haenel wrote: > Hi, > > when i look at the documentation for NiftiDataset, i see that samples has a > default of None. This seems to suggest i could use: > > NiftiDataset() > > to instantiate an empty dataset, which however raises a ValueError > > My use case is such that i want to instantiate an empty dataset and then use > "+=" to successively populate the dataset in a loop.
I see no easy solution to this problem. The presence or absence of the samples argument is the trigger for the 'copy-constructor-behavior' of the dataset classes -- a feature that is heavily used in PyMVPA. Since Python cannot handle more than one contructor it has been implemented this way. It might be worth to evaluate that carefully and see whether such change is doable. We might achieve that during the upcoming refactoring of the dataset base class -- will keep it in mind... Thanks, Michael -- GPG key: 1024D/3144BE0F Michael Hanke http://apsy.gse.uni-magdeburg.de/hanke ICQ: 48230050 _______________________________________________ Pkg-ExpPsy-PyMVPA mailing list [email protected] http://lists.alioth.debian.org/mailman/listinfo/pkg-exppsy-pymvpa
