WIP: Adjustments for common Table interface

[andreasnoack]

This is a tiny one but there might be more changes that I'd like to propose in order to get a common interface for tables that StatsModels can use. The change here is consistent with the behavior in NullableArrays and it might help in the construction of contrasts. I don't think there should be any NullableArrays specific code in StatsModels and the price would then have to be that the ContrastsMatrix has Nullables in it but I guess that is in line with the idea of propagating nullables.

• Return Array{Nullable{T}} instead of NullableArray{T} in unique

[codecov-io]

Codecov Report

Merging #55 into master will not change coverage.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master      #55   +/-   ##
=======================================
  Coverage   90.89%   90.89%           
=======================================
  Files           9        9           
  Lines         494      494           
=======================================
  Hits          449      449           
  Misses         45       45

Impacted Files	Coverage Δ
src/nullablearray.jl	71.79% <100%> (ø)	✅

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 388575a...69ca30f. Read the comment docs.

[coveralls]

Coverage remained the same at 90.891% when pulling 69ca30f on andreasnoack:master into 388575a on JuliaData:master.

[nalimilan]

Why not. Performance isn't typically an issue with such short vectors.

There's another occurrence to remove in src/subarray.jl. Then the only remaining use of NullableArrays (except for tests, which is OK) is for the optimized version of fill_refs!, used by cut. I guess it can be good enough for now to use the generic AbstractArray{<:Nullable} API, which shouldn't be that slow, and remove the specialized version and the dependency on NullableArray altogether.

[nalimilan]

This has been fixed by the move to Nulls.