No, I didn't test, I was just reading the code, but I hadn't checked for all occurrences of K_COUNT, so I hadn't noticed that it also occurs in the allowedFunctionName grammar production rule. And I found the code that dynamically creates a count function for each type here: https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/cql3/functions/AggregateFcts.java#L885
So, sorry for the confusion. Some comments in the grammar would have helped. And it's not in the DataStax doc, but that's another issue I can follow up on. -- Jack Krupansky On Mon, Apr 18, 2016 at 10:12 AM, Benjamin Lerer < benjamin.le...@datastax.com> wrote: > May be I misunderstood you. > Do you mean that you tested it and that it is not working on the version > you used? > > On Mon, Apr 18, 2016 at 3:50 PM, Jack Krupansky <jack.krupan...@gmail.com> > wrote: > > > The CQL spec for COUNT says: > > > > "It also can be used to count the non null value of a given column. > > Example: > > > > SELECT COUNT(scores) FROM plays;" > > > > But, the parser only recognizes COUNT(*) and COUNT(1). > > > > See: > > https://cassandra.apache.org/doc/cql3/CQL-3.0.html > > https://github.com/apache/cassandra/blob/trunk/src/antlr/Parser.g#L276 > > > > Does this need a Jira, or can somebody just fix it to avoid the > paperwork? > > > > Or, was this actually supposed to work and the bug is some missing > > implementation? > > > > Thanks. > > > > -- Jack Krupansky > > >
