Support _DF (or something) for full DataFrame access

[tshort]

We used to have this "hidden" feature where you could use _DF to access the whole DataFrame (or SubDataFrame in a group operation). It is occasionally nice. I think it got coded out in #73.

[piever]

I've been playing with this in JuliaDBMeta and I think the nicest is to use _ (just like what Query does, as well as the proposed new anonymous function syntax). It could actually be a non-hidden feature and make the _I_ syntax less necessary as one can just do _[expr] which is even shorter.

[tshort]

Interesting on the JuliaDBMeta approach. I'm curious if you run into stumbling blocks with _. BTW, great to see JuliaDBMeta. I'm looking forward to using JuliaDB more, and having JuliaDBMeta will make it nicer!

[piever]

_ works pretty well as it is warned against in Julia's normal syntax so it doesn't really appear in normal code.
For now I'm making _ be the whole DataFrame for "column based macros" and the whole row as a NamedTuple for "row based macros". Ideally it should also be the appropriate subdataframe when working with grouping data but I haven't figured out the grouping part yet (JuliaDB grouping API is already quite powerful, it's not too obvious how to improve on it...).

[pdeffebach]

Closed in favor of #229