I found it in https://github.com/antlr/grammars-v4.git and I suspected it was wrong. I assume it should be that the tuple can contain 1 or more elements and the elements may be of type tuple, constant, map, list or set.
Does that make sense? I think that is what I saw in the code base. On Thu, 10 Mar 2022 at 09:38, Benjamin Lerer <ble...@apache.org> wrote: > Hi Claude, > > I am not aware of the CqlParser.g4 file in our code base. Where did you > find that file? > > At first glance effectively something looks wrong in the syntax. The > construct ((4 ,5 ), 6, (7, 8)) should be legal in CQL. > > Le jeu. 10 mars 2022 à 06:50, Claude Warren <claude.war...@instaclustr.com> > a écrit : > >> I have been looking at CqlParser.g4 file for cql3 and have a question >> about assignment tuples. The assignment tuple is defined as : >> >> assignmentTuple >> : syntaxBracketLr ( >> constant ((syntaxComma constant)* | (syntaxComma assignmentTuple)*) >> | >> assignmentTuple (syntaxComma assignmentTuple)* >> ) syntaxBracketRr >> ; >> >> which I read to be ( constant [, constant | tuple ... ]) or ( tuple [, >> tuple...]) . So the construct ((4 ,5 ), 6, (7, 8)) is not a legal tuple.2 >> questions: >> >> 1. Is my interpretation of the grammar correct? >> 2. Is my example tuple supposed to be allowed? >> >> >> Claude >> >> -- >> >> [image: Instaclustr logo] >> >> >> *Claude Warren* >> >> Principal Software Engineer >> >> Instaclustr >> >> >> >>
