Le ven. 16 août 2019 à 17:25, Alex Herbert <alex.d.herb...@gmail.com> a écrit : > > > > > On 16 Aug 2019, at 16:06, Gilles Sadowski <gillese...@gmail.com> wrote: > > > > Hello. > > > > Le jeu. 15 août 2019 à 11:33, akash srivastava <akashsp...@gmail.com> a > > écrit : > >> > >> Here is the link for the related bug: > >> https://issues.apache.org/jira/browse/MATH-1495 > >> > >> I have suggested two possible fixes when NaturalRanking#rank() is called on > >> an array of all NaNs: > >> 1) throwing an IllegalArgumentException > >> 2) Returning an empty array > >> > >> I want to create a pull request regarding it, which fix do you suggest I > >> should go for? > > > > There are rationale for each. An exception would catch a programming bug > > early. > > But returning an empty array would allow concise code when such a case is > > not a > > bug. > > WDYT? Are there use-cases? > > IIUC this is for the case when the NaNStrategy is REMOVED. So the result > should match what happens when you call NaturalRanking.rank() with an empty > array.
+1 > Is that an empty array result or an exception? We could nevertheless change the behaviour (consistently) if one or the other laternative is deemed better. Regards, Gilles --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org