Comment #1 on issue 907 by jochen.voss: add QRsolve to matrices
http://code.google.com/p/sympy/issues/detail?id=907
Here is a patch to implement the missing method (already posted to
sympy-patches).
Attachments:
0001-Add-a-new-Matrix.QRsolve-method.patch 3.2 KB
--
You received this
Status: Accepted
Owner: jochen.voss
Labels: Type-Defect Priority-Medium
New issue 1421 by jochen.voss: QRdecomposition fails for some matrices
http://code.google.com/p/sympy/issues/detail?id=1421
Hello,
on the current git version of sympy I got a strange failure of the
Matrix.QRdecomposition
Comment #1 on issue 1421 by ondrej.certik: QRdecomposition fails for some
matrices
http://code.google.com/p/sympy/issues/detail?id=1421
Thanks for the bug report. One should figure out why the assertion fails:
-- 968 assert Q[:,j].norm() == 1
e.g. what the norm() is, etc.
--
Comment #1 on issue 1416 by jochen.voss: integrate((x**3 - x)/ (x - 1), x)
gives traceback
http://code.google.com/p/sympy/issues/detail?id=1416
Mateusz's patch at
http://groups.google.com/group/sympy-patches/browse_thread/thread/5be1e851b4568925
seems to fix the issue. With his patch applied
Comment #1 on issue 1417 by jochen.voss: integrate(2**x - 2*x, x) gives
traceback
http://code.google.com/p/sympy/issues/detail?id=1417
Mateusz's patch at
http://groups.google.com/group/sympy-patches/browse_thread/thread/5be1e851b4568925
seems to fix the issue. With his patch applied I get:
Comment #6 on issue 1419 by jochen.voss: integrate(x/(x**2-9), x) gives
wrong answers
http://code.google.com/p/sympy/issues/detail?id=1419
Mateusz's patch at
http://groups.google.com/group/sympy-patches/browse_thread/thread/5be1e851b4568925
seems to fix this issue. With his patch applied I
Comment #2 on issue 1421 by jochen.voss: QRdecomposition fails for some
matrices
http://code.google.com/p/sympy/issues/detail?id=1421
Actully the three lines from the traceback, now that I look at them, show
what's
going on:
966 R[j,j] = tmp.norm()
967
Comment #3 on issue 1421 by jochen.voss: QRdecomposition fails for some
matrices
http://code.google.com/p/sympy/issues/detail?id=1421
By the way, the algorithm looks like Gram-Schmidt orthonormalisation to
me. Is this
right? If so, the current implementation of QRsolve would probably not