gnu-octave
deprecation warning using `solve`: "non-Expr objects in a Matrix is deprecated"
When trying the following MATLAB example code:
syms u v
eqns = [2*u^2 + v^2 == 0, u - v == 1];
vars = [v u];
[solv, solu] = solve(eqns,vars)
we get a deprecation warning, the output is:
Symbolic pkg v3.2.1: Python communication link active, SymPy v1.13.3.
<stdin>:7: SymPyDeprecationWarning:
non-Expr objects in a Matrix is deprecated. Matrix represents
a mathematical matrix. To represent a container of non-numeric
entities, Use a list of lists, TableForm, NumPy array, or some
other data structure instead.
See https://docs.sympy.org/latest/explanation/active-deprecations.html#deprecated-non-expr-in-matrix
for details.
This has been deprecated since SymPy version 1.9. It
will be removed in a future version of SymPy.
solv = (sym 2×1 matrix)
⎡ 2 √2⋅ⅈ⎤
⎢- ─ + ────⎥
⎢ 3 3 ⎥
⎢ ⎥
⎢ 2 √2⋅ⅈ⎥
⎢- ─ - ────⎥
⎣ 3 3 ⎦
solu = (sym 2×1 matrix)
⎡1 √2⋅ⅈ⎤
⎢─ + ────⎥
⎢3 3 ⎥
⎢ ⎥
⎢1 √2⋅ⅈ⎥
⎢─ - ────⎥
⎣3 3 ⎦
If there is another way to solve a linear system without getting the deprecation then I'd be happy too - but I can't find any.
Hi @cbm755 and @CarloWood,
Regarding the deprecation warning for this issue, I ran the MATLAB example code with the following versions:
- Octave: 9.2.0
- SymPy: 1.10.1
- Symbolic package: 3.2.1
Here’s the code I executed:
octave:1> pkg load symbolic
octave:2> version
ans = 9.2.0
octave:3> syms u v
Symbolic pkg v3.2.1: Python communication link active, SymPy v1.10.1.
octave:4> eqns = [2*u^2 + v^2 == 0, u - v == 1];
octave:5> vars = [v u];
octave:6> [solv, solu] = solve(eqns,vars)
solv = (sym 2x1 matrix)
[ ___ ]
[1 \/ 2 *I]
[- + -------]
[3 3 ]
[ ]
[ ___ ]
[1 \/ 2 *I]
[- - -------]
[3 3 ]
solu = (sym 2x1 matrix)
[ ___ ]
[ 2 \/ 2 *I]
[- - + -------]
[ 3 3 ]
[ ]
[ ___ ]
[ 2 \/ 2 *I]
[- - - -------]
[ 3 3 ]
Even after enabling all warnings with warning('on', 'all'), I did not encounter any deprecation warnings. However, I did see several other warnings related to boolean values and string conversions:
octave:7> warning('on', 'all')
octave:8> [solv, solu] = solve(eqns,vars)
warning: Using an object of size 3x1 as a boolean value implies all().
warning: called from
compare_versions at line 115 column 7
python_ipc_driver at line 36 column 9
pycall_sympy__ at line 164 column 11
solve at line 197 column 9
warning: Using an object of size 2x1 as a boolean value implies all().
warning: called from
compare_versions at line 119 column 7
python_ipc_driver at line 36 column 9
pycall_sympy__ at line 164 column 11
solve at line 197 column 9
warning: implicit conversion from string to real N-D array
warning: called from
compare_versions at line 159 column 8
python_ipc_driver at line 36 column 9
pycall_sympy__ at line 164 column 11
solve at line 197 column 9
warning: implicit conversion from string to real N-D array
warning: called from
compare_versions at line 159 column 8
python_ipc_driver at line 36 column 9
pycall_sympy__ at line 164 column 11
solve at line 197 column 9
warning: concatenation of different character string types may have unintended consequences
warning: called from
strjoin at line 84 column 10
write_lines at line 28 column 3
python_ipc_popen2 at line 145 column 3
python_ipc_driver at line 62 column 15
pycall_sympy__ at line 164 column 11
solve at line 197 column 9
warning: concatenation of different character string types may have unintended consequences
warning: called from
strjoin at line 84 column 10
write_lines at line 28 column 3
python_ipc_popen2 at line 180 column 3
python_ipc_driver at line 62 column 15
pycall_sympy__ at line 164 column 11
solve at line 197 column 9
warning: concatenation of different character string types may have unintended consequences
warning: called from
strjoin at line 84 column 10
write_lines at line 28 column 3
python_ipc_popen2 at line 182 column 3
python_ipc_driver at line 62 column 15
pycall_sympy__ at line 164 column 11
solve at line 197 column 9
solv = (sym 2x1 matrix)
[ ___ ]
[1 \/ 2 *I]
[- + -------]
[3 3 ]
[ ]
[ ___ ]
[1 \/ 2 *I]
[- - -------]
[3 3 ]
solu = (sym 2x1 matrix)
[ ___ ]
[ 2 \/ 2 *I]
[- - + -------]
[ 3 3 ]
[ ]
[ ___ ]
[ 2 \/ 2 *I]
[- - - -------]
[ 3 3 ]
Next, I tested with the following versions:
- Octave: 9.2.0
- SymPy: 1.13.3
- Symbolic package: 3.2.1
Again, I ran the same code:
octave:3> clear all;
octave:4> version
ans = 9.2.0
octave:5> pkg load symbolic
octave:6> syms u v
Symbolic pkg v3.2.1: Python communication link active, SymPy v1.13.3.
octave:7> eqns = [2*u^2 + v^2 == 0, u - v == 1];
octave:8> vars = [v u];
octave:9> [solv, solu] = solve(eqns,vars)
solv = (sym 2x1 matrix)
[ ___ ]
[ 2 \/ 2 *I]
[- - - -------]
[ 3 3 ]
[ ]
[ ___ ]
[ 2 \/ 2 *I]
[- - + -------]
[ 3 3 ]
solu = (sym 2x1 matrix)
[ ___ ]
[1 \/ 2 *I]
[- - -------]
[3 3 ]
[ ]
[ ___ ]
[1 \/ 2 *I]
[- + -------]
[3 3 ]
Again, no deprecation warnings appeared after enabling all warnings.
Additional Note on Using vpasolve:
If you're looking for an alternative to solve, consider using the vpasolve function, which requires an initial guess for each variable. Here’s the basic syntax:
syms x1 x2 x3 ... xn
eqns = [eqn1; eqn2; eqn3; ...; eqnn];
initial_guess = [g1; g2; g3; ...; gn];
vpasolve(eqns, [x1; x2; x3; ...; xn], initial_guess)
Here’s an example with my previous equations:
octave:21> clear all;
octave:22> version
ans = 9.2.0
octave:23> pkg load symbolic
octave:24> syms u v
Symbolic pkg v3.2.1: Python communication link active, SymPy v1.13.3.
octave:25> eqns = [2*u^2 + v^2 == 0; u - v == 1]
eqns = (sym 2x1 matrix)
[ 2 2 ]
[2*u + v = 0]
[ ]
[ u - v = 1 ]
octave:26> vpasolve(eqns, [u; v], [1i; 1i])
ans = (sym 2x1 matrix)
[0.33333333333333333333333333333333 + 0.47140452079103168293389624140323*I ]
[ ]
[-0.66666666666666666666666666666667 + 0.47140452079103168293389624140323*I]
octave:27> vpasolve(eqns, [u; v], [4i; 4i])
ans = (sym 2x1 matrix)
[0.33333333333333333333333333333333 + 0.47140452079103168293389624140323*I ]
[ ]
[-0.66666666666666666666666666666667 + 0.47140452079103168293389624140323*I]
octave:28> vpasolve(eqns, [u; v], [-4i; -4i])
ans = (sym 2x1 matrix)
[0.33333333333333333333333333333333 - 0.47140452079103168293389624140323*I ]
[ ]
[-0.66666666666666666666666666666667 - 0.47140452079103168293389624140323*I]
Thank you for your attention to this matter. I appreciate any advice, corrections, or guidance you can provide moving forward.
Hi @cbm755 , Could you kindly review the comment on this issue when you have a chance? Thank you for your time and any guidance you can provide.
This looks like a different bug:
warning: Using an object of size 3x1 as a boolean value implies all().
can you file a different issue about it?
@Sonu0305, maybe one issue to cover things about by warning('on', 'all')?
The reason (I think) that this is unrelated to the Non-Expr warning is that the Non-Expr warning comes from Python (SymPy) but the Octave warning command controls Octave's warnings.