At this point, backward compatibility. Enough people use it expecting it to work that it would be bad to change the behavior.
On Mon, May 9, 2016 at 6:42 PM, dandl <david at andl.org> wrote: > > bounces at mailinglists.sqlite.org] On Behalf Of Keith Medcalf > > > Why are you using BOTH distinct and group by on the same column? You > only > > need one or the other. If you are redundantly redundant I would hope > that > > the optimizer makes redundant (as in gets rid of, for those that are not > > English) the redundancies ... > > This is generated code. Since Andl does not allow any duplicate rows, every > SELECT gets a DISTINCT unless the query provably cannot generate > duplicates. > You need both GROUP BY and DISTINCT in cases where there is an aggregate > function (and some others). Say: > > SELECT DISTINCT SUM(X) AS Y FROM T GROUP BY Z; > > There is no way to predict from the query how many rows this will generate. > Without DISTINCT it can generate duplicates. > > My question was really about why Sqlite did not complain on what is > actually > not a valid query. [Andl is still a work in progress.] > > Regards > David M Bennett FACS > > Andl - A New Database Language - andl.org > > > > > > _______________________________________________ > sqlite-users mailing list > sqlite-users at mailinglists.sqlite.org > http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users > -- Scott Robison
