Simplify and increase robustness of sorting

[LilithHafner]

This way, if more keyword arguments are added, they will automatically be supported

Seeing these passthrough methods makes me think it might be worth making DataSetColumn <: AbstractVector, but I'm not familiar enough with this package to know if that is a good idea.

[sl-solution]

Because we want to keep track of datasets' columns, we have not set DatasetColumn as a subtype of AbstractVector. We like to track columns to make sure that if a column is changed by something rather than IMD's APIs, its parent's metadata is updated (for example if a dataset is sorted by the modified column, then its sorting status must be removed).

[codecov-commenter]

Codecov Report

Merging #95 (2e3e6a5) into master (44445f6) will increase coverage by 0.05%.
The diff coverage is 85.45%.

@@            Coverage Diff             @@
##           master      #95      +/-   ##
==========================================
+ Coverage   60.89%   60.95%   +0.05%     
==========================================
  Files          51       51              
  Lines       11241    11245       +4     
==========================================
+ Hits         6845     6854       +9     
+ Misses       4396     4391       -5     

Impacted Files	Coverage Δ
src/abstractdataset/dscol.jl	14.06% <0.00%> (+0.32%)	⬆️
src/sort/util.jl	48.68% <55.55%> (ø)
src/dataset/constructor.jl	98.26% <100.00%> (ø)
src/dataset/modify.jl	45.56% <100.00%> (ø)
src/sort/qsort.jl	98.49% <100.00%> (+0.05%)	⬆️
src/sort/sort.jl	78.77% <100.00%> (ø)
src/join/join_dict.jl	86.02% <0.00%> (+0.43%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more