Robert Bradshaw wrote: > On Apr 7, 2008, at 10:07 AM, Jason Grout wrote: >> In trying to unify the linear algebra syntax, I came across the >> following behavior: >> >> >> sage: a=matrix(2,range(4)) >> sage: a.nrows() >> 2 >> sage: a._nrows >> ---------------------------------------------------------------------- >> ----- >> <type 'exceptions.AttributeError'> Traceback (most recent >> call last) >> >> /home/grout/sage/devel/sage-main/sage/misc/<ipython console> in >> <module>() >> >> <type 'exceptions.AttributeError'>: >> 'sage.matrix.matrix_integer_dense.Matrix_integer_de' object has no >> attribute '_nrows' >> sage: a.__dict__ >> ---------------------------------------------------------------------- >> ----- >> <type 'exceptions.AttributeError'> Traceback (most recent >> call last) >> >> /home/grout/sage/devel/sage-main/sage/misc/<ipython console> in >> <module>() >> >> <type 'exceptions.AttributeError'>: >> 'sage.matrix.matrix_integer_dense.Matrix_integer_de' object has no >> attribute '__dict__' >> >> >> I noted that the nrows() function just returned self._nrows. So how >> come I can't access _nrows? Does it have to do with Cython and >> inheritance? > > Yes, this is a Cython vs. Python issue. a._nrows is literally a c int > in the object structure, not a member of a dictionary. In fact, > Cython classes don't even have dictionaries. From Cython a._nrows is > probably on the order of two or three magnitudes faster than a.nrows() > > We could declare _nrows and _ncols to be readonly, in which case a > __getattr__ (or more technically a property accessor method) method > would be created for them.
Thanks for the explanation. The issue came up when I wanted to just do M=self.transpose() or M=self, depending on if the user wanted right or left action computations (e.g., right vs. left kernels). However, then calls to M._nrows (which used to be calls to self._nrows) don't work. I'll just set an M_nrows variable to either self._nrows or self._ncols, whichever is needed. That way we still have the order-of-magnitude speedup. Unless someone has a better idea. Thanks, Jason --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
