I think the name should be something like "solve2" and provide a couple year deprecation warning after which "solve" points to "solve2".
Jonathan On Friday, March 4, 2022 at 1:40:58 AM UTC-6 [email protected] wrote: > I think changing this will break tons of code in the wild. Isnt it best > make a new "solve_new" and then leave solve be (maybe with a deprecation > warning. You could call it `solve_equations` or something. > > Jason > moorepants.info > +01 530-601-9791 <(530)%20601-9791> > > > On Thu, Mar 3, 2022 at 11:45 PM Oscar Benjamin <[email protected]> > wrote: > >> On Thu, 3 Mar 2022 at 20:28, Aaron Meurer <[email protected]> wrote: >> > On Thu, Mar 3, 2022 at 8:28 AM Chris Smith <[email protected]> wrote: >> > > >> > > Although the dict=True or set=True will give a standard output, can >> we at least unify the case for when variables are given so we always get a >> list of one or more dictionaries? So the above would be `[{x: -sqrt(y)}, >> {x: sqrt(y)}]` and `[{x: y}]`, respectively. This would then make `solve` >> always give a list of a) values for a univariate expression, b) a list of >> one or more dictionaries for every other case. (Case (a) will give a list >> of dictionaries if `dict=True`.) >> > >> > Changing the output type could break code that solves a specific >> > equation. I am doubtful whether any users actually understand the >> > output type behavior of solve without the dict=True flag. So >> > personally I think we should clean it up. We already recommend using >> > dict=True to get consistent output types, and this would only affect >> > users who aren't doing that. >> >> It would be much better if solve always worked like that but obviously >> that's not a backwards compatible change and I'm not sure what would >> break. I don't think that the type of the output should depend on the >> number of solutions although it could depend on the type of the >> arguments (e.g a single equation vs a list of equations). At the >> moment the type of the output can depend on the equations themselves >> which is wildly awkward: >> >> In [3]: solve([x-1], [x]) >> Out[3]: {x: 1} >> >> In [4]: solve([x**2-1], [x]) >> Out[4]: [(-1,), (1,)] >> >> The ideal output for solve is absolutely a list of dicts though and it >> would be good to make that the default at least when given a list of >> equations. I guarantee that a lot of notebooks or other little bits of >> code will depend on each of these random special cases though. >> >> -- >> Oscar >> >> -- >> You received this message because you are subscribed to the Google Groups >> "sympy" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> > To view this discussion on the web visit >> https://groups.google.com/d/msgid/sympy/CAHVvXxROqs0CpGmSX2%3Dk8xJgTos9FPpP1x0GYP0aHCdiYErDnA%40mail.gmail.com >> . >> > -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/d837d55d-2def-4b25-812f-ea22aa89d017n%40googlegroups.com.
