David Sisson

There is a class of essentially recursive plans that we don't have an explicit definition for at the moment but probably should. These are Common Table Expressions (CTEs). As such...

I agree that we do need to know and describe those limits as well as the overall graph structure. If it matters to you we can try to tackle those...

This behavior is identical to what happens with fixed_char, var_char, and decimal literals. The way this is addressed in Isthmus is to immediately cast the literal to the appropriate precision....

The core issue is that there isn't a fixed issue of the literal (it's either microseconds or nanoseconds but not both). I'd argue that the definition should just be nanoseconds...

I'm onboard with fully specifying the literal. We should also include the missing timezone part in the tz variant as well.

Similar functionality will be added in #528 so I believe we can close this PR.

Ah, yes, thanks for the correction!

@vibhatha It seems like we aren't going to pursue this option. Should we close?

@westonpace @mbrobbel Are we good to merge this now?

This was discussed in the Oct 11, 2013 community meeting (as well as the April 23, 2013 community meeting). The plan there is to instead use the same language used...