Unify argument conventions

[jseabold]

I know how hard it can be to stick to conventions, but it might be nice to try to unify arguments that do similar things. For example, the argument to select/name columns is not consistent. I realize it's a bit difficult since columns usually means labels and there's an axis argument. I don't know to what extent it's feasible because of this, but maybe all of these should be column(s)? Or even name(s) since this doesn't imply an axis but can be inferred from context when there's no axis argument. The few that come to mind since I just used them.

dropna -- subset
drop_duplicates -- cols
DataFrame -- columns
Series -- name

Just a thought.

[joeb1415]

+1.
DataFrame.pivot_table() takes "cols" as an argument. "columns" is more common. Motion to change pivot_table's argument to "columns" instead?

[joeb1415]

While we're at it with pivot_table, could we accept "index" instead of / in addition to "rows" as a parameter name?

[jreback]

this was fixed by #5505, #6680

[jseabold]

Fantastic. Thanks!