I can't figure out how to create a new sparse matrix (see below).

Sorry,
Dave

M = Matrix_integer_sparse.__new__(Matrix_integer_sparse,
sage.matrix.matrix_space.MatrixSpace(ZZ, self._nrows-1, self._ncols,
sparse=True), None, None, None)


-------------------
...
    def _delete_row_(self, Py_ssize_t row_number):
        cdef Py_ssize_t i
        cdef mpz_vector* self_row, *M_row
        cdef Matrix_integer_sparse M

        M = Matrix_integer_sparse.__new__(Matrix_integer_sparse,
sage.matrix.matrix_space.MatrixSpace(ZZ, self._nrows-1, self._ncols,
sparse=True), None, None, None)
 
^
------------------------------------------------------------

/usr/local/sage-devel/devel/sage-devel/sage/matrix/
matrix_integer_sparse.pyx:221:101: Compiler crash in
AnalyseExpressionsTransform

ModuleNode.body = StatListNode(matrix_integer_sparse.pyx:23:0)
StatListNode.stats[13] = StatListNode(matrix_integer_sparse.pyx:43:5)
StatListNode.stats[0] = CClassDefNode(matrix_integer_sparse.pyx:43:5,
    as_name = u'Matrix_integer_sparse',
    base_class_module = u'matrix_sparse',
    base_class_name = u'Matrix_sparse',
    class_name = u'Matrix_integer_sparse',
    module_name = '',
    visibility = 'private')
CClassDefNode.body = StatListNode(matrix_integer_sparse.pyx:55:4)
StatListNode.stats[9] = DefNode(matrix_integer_sparse.pyx:216:4,
    name = u'_delete_row_',
    num_required_args = 2,
    reqd_kw_flags_cname = '0')
File 'Nodes.py', line 360, in analyse_expressions: StatListNode
(matrix_integer_sparse.pyx:217:8)
File 'Nodes.py', line 2835, in analyse_expressions:
SingleAssignmentNode(matrix_integer_sparse.pyx:221:41)
File 'Nodes.py', line 2928, in analyse_types: SingleAssignmentNode
(matrix_integer_sparse.pyx:221:41)
File 'ExprNodes.py', line 2281, in analyse_types: SimpleCallNode
(matrix_integer_sparse.pyx:221:41)
File 'ExprNodes.py', line 3105, in analyse_types: TupleNode
(matrix_integer_sparse.pyx:221:41,
    is_sequence_constructor = 1)
File 'ExprNodes.py', line 2972, in analyse_types: TupleNode
(matrix_integer_sparse.pyx:221:41,
    is_sequence_constructor = 1)
File 'ExprNodes.py', line 2522, in analyse_types: GeneralCallNode
(matrix_integer_sparse.pyx:221:101)

Compiler crash traceback from this point on:
  File "/usr/local/sage-devel/local/lib/python2.5/site-packages/Cython/
Compiler/ExprNodes.py", line 2532, in analyse_types
    if hasattr(self.function, 'entry') and not
self.function.entry.as_variable:
AttributeError: 'NoneType' object has no attribute 'as_variable'
Error running command, failed with status 256.
sage: There was an error installing modified sage library code.


On Jun 2, 7:28 pm, William Stein <wst...@gmail.com> wrote:
> On Tue, Jun 2, 2009 at 7:22 PM, davidp <dav...@reed.edu> wrote:
>
> > I tried adding a delete_row method to matrix_integer_sparse.pyx but
> > stopped after getting the error message:
>
> Just def your method -- do not cpdef it.  You can still use Cython
> code in a cpdef's method and it will be just as fast.
>
> Make sure that your "delete row" method returns a *new* matrix by the way...
>
> William
>
>
>
>
>
> > +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> > Error converting Pyrex file to C:
> > ------------------------------------------------------------
> > ...
> >            add_mpz_vector_init(&M._matrix[i], &self._matrix[i], &
> > (<Matrix_integer_sparse>right)._matrix[i], mul)
> >        mpz_clear(mul)
> >        return M
>
> > # added by David Perkinson
> >    cpdef ModuleElement _delete_row_(self, Py_ssize_t row_number):
> >         ^
> > ------------------------------------------------------------
>
> > /usr/local/sage-devel/devel/sage-devel/sage/matrix/
> > matrix_integer_sparse.pyx:216:10: C method '_delete_row_' not
> > previously declared in definition part of extension type
> > Error running command, failed with status 256.
> > sage: There was an error installing modified sage library code.
>
> > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>
> > I might have to actually read the Cython documentation!
>
> > Dave
>
> > On Jun 2, 5:43 pm, Jason Grout <jason-s...@creativetrax.com> wrote:
> >> William Stein wrote:
> >> > On Mon, Jun 1, 2009 at 5:16 PM, davidp <dav...@reed.edu> wrote:
> >> >> Is there a fast way to create a submatrix?
>
> >> >> -------------------------------------------------
>
> >> >> sage: version()
> >> >> 'Sage Version 4.0.alpha0, Release Date: 2009-05-15'
> >> >> sage: G = graphs.GridGraph([100,100])
> >> >> sage: L = G.laplacian_matrix()
> >> >> sage: L
> >> >> 10000 x 10000 sparse matrix over Integer Ring
> >> >> sage: time M = L[1:9999,1:9999]
> >> >> CPU times: user 24.93 s, sys: 0.04 s, total: 24.97 s
> >> >> Wall time: 25.27 s
>
> >> >> ----------------------------------------------------
>
> >> >> I am just interested in deleting a single row and column of the matrix
> >> >> (not necessarily the first).
>
> >> >> Thanks,
>
> >> > There is no fast way to do that right now.  One could easily add code
> >> > to SAGE_ROOT/devel/sage/sage/matrix/matrix_integer_sparse.pyx that
> >> > would provide blazingly fast deletion of a row, and reasonably fast
> >> > deletion of a column.   Of course it would be better to implement
> >> > arbitrary slicing in some optimized way in matrix_integer_sparse.pyx.
> >> > I hope somebody does so.
>
> >> Just looking at the generic code, it seems that it goes through each and
> >> every index position in the slice, setting the new matrix entry to the
> >> old one.  This is obviously the wrong thing to do for sparse matrices,
> >> and can likely trivially be made faster.  I think all you may have to do
> >> is override the matrix_from_rows_and_columns for sparse matrices.
>
> >> Jason
>
> --
> William Stein
> Associate Professor of Mathematics
> University of Washingtonhttp://wstein.org
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-support@googlegroups.com
To unsubscribe from this group, send email to 
sage-support-unsubscr...@googlegroups.com
For more options, visit this group at 
http://groups.google.com/group/sage-support
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---

Reply via email to