Hello all, In process of working on https://issues.apache.org/jira/browse/GRIFFIN-205 I'v implemented a patch for "batch" dq type. Now I see at least 2 options: - Apply the patch as is and create separate ticket for "streaming" dq type. As an advantage here we will improve "batch" dq type as soon as possible (and it covers our needs for now). A disadvantage is an inconsistent functionality provided by "batch" and "streaming" types. - Implement the same logic for streaming and only then merge it to master. Need to keep in mind, that for batch case I've added automated tests which cover basic cases. Looks like for "streaming" it will be much more difficult. It requires more changes in code (due to some abstractions leaks in streaming processing) and as a result it requires quite good testing, but we don't use this mode in our cases (yet).
My question is how strong is Apache Griffin's features consistency regarding "batch" and "streaming" dq types? I would go with the first option if you don't mind. Objections? Thanks, Artem.
