Circshift bug for sparse logical.

Question

Jan Orwat 2015-12-8

3
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/259365-circshift-bug-for-sparse-logical

编辑： James Tursa 2024-10-23，21:15

Hello, does anybody know if it has been reported?

Circshift seems to shift sparse logicals properly, but changes trues to falses, which leads to several logical problems:

 >> C = circshift(sparse(true),1)
 C =
   (1,1)        0
 >> C == ~C
 ans =
   (1,1)        1
 >> C&C
 ans =
   (1,1)        0
 >> C|C
 ans =
   (1,1)        1
 % etc.

It's not a major problem since sparse logicals doesn't seem to be in common use and workaround with indexing is very simple. Bug confirmed in 2015a and 2015b, denied in 2010b, 2013a and 2013b. It probably exists since last changes in circshift in 2014a.

source

3 个评论
显示 1更早的评论隐藏 1更早的评论

DGM 2024-10-23，0:59

在 MATLAB Online 中打开

Welp, I guess that was changed somewhere between R2017b and R2019b. I didn't see anything in bug reports, but ... yeah.

C = circshift(sparse(true),1)
C = sparse logical
   (1,1)      1
C == ~C
ans = sparse logical
   All zero
C&C
ans = sparse logical
   (1,1)      1
C|C
ans = sparse logical
   (1,1)      1

James Tursa 2024-10-23，21:01

编辑：James Tursa 2024-10-23，21:15

在 MATLAB Online 中打开

Looks like bug is fixed, at least for this example (using sarek sparse matrix checker from FEX):

https://www.mathworks.com/matlabcentral/fileexchange/158161-sarek-sparse-analyzer-real-et-komplex?s_tid=srchtitle

>> st = sparse(true)
st =
sparse logical
(1,1)      1
>> sarek(st)
SAREK -- Sparse Analyzer Real Et Komplex , by James Tursa
Compiled in version R2020a (with -R2018a option)
Running  in version R2020a
Matrix is logical ...
    M = 1 >= 0 OK ...
    N = 1 >= 0 OK ...
    Nzmax = 1 >= 1 OK ...
    Jc[0] == 0 OK ...
    Jc[N] = 1 <= Nzmax = 1 OK ...
    Jc array OK ...
    Ir array OK ...
    All stored elements nonzero OK ...
    
All sparse integrity checks OK
ans =
0
>> 
>> cst = circshift(st,1)
cst =
sparse logical
(1,1)      1
>> sarek(cst)
SAREK -- Sparse Analyzer Real Et Komplex , by James Tursa
Compiled in version R2020a (with -R2018a option)
Running  in version R2020a
Matrix is logical ...
    M = 1 >= 0 OK ...
    N = 1 >= 0 OK ...
    Nzmax = 1 >= 1 OK ...
    Jc[0] == 0 OK ...
    Jc[N] = 1 <= Nzmax = 1 OK ...
    Jc array OK ...
    Ir array OK ...
    All stored elements nonzero OK ...
    
All sparse integrity checks OK
ans =
0