drill-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jacques Nadeau <jacq...@dremio.com>
Subject Re: ensureAtLeastOneField: obsolete? needed?
Date Mon, 05 Oct 2015 16:13:27 GMT
On Sun, Oct 4, 2015 at 8:56 PM, Daniel Barclay <dbarclay@maprtech.com>

> In what sense is that not supported?  For example, where would returning
> zero columns be expected to break?

The system wasn't designed with this in mind. That is why the iterator
validator ensures that this isn't happening. If you want to add this as a
new supported behavior, I think you would need to create a set of tests
around each of the operators ensuring this behavior works appropriately
(before removing the iterator validation). I'm not sure how good the test
coverage is for this scenario so I don't think running the units tests with
this fundamental change is satisfactory for determining it is okay.

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message