I think the syntax is a bit confusing. It mixes up SQL-like `GROUP BY` with 
Beam's `GroupByKey`, so for me it takes a second to figure out what this 
translates to. E.g. if I read this coming from SQL, it's unclear why I am 
grouping by `userId` and `country` but then aggregating only `cost`, but not 
grouping by `cost`.

It's also unclear what the output schema here is. Is it only the fields in the 
`Group.`? How does it work with `Select`?

Would it be better to follow SQL-like approach? I.e. add aggregations in 
`Select` and use `Group` to only specify the keys?

[ Full content available at: https://github.com/apache/beam/pull/6298 ]
This message was relayed via gitbox.apache.org for [email protected]

Reply via email to