REF: Unify creation of dataframe for C and Python parser engines

[lithomas1]

We should unify the process of creating the dataframe from index,names,data for the C and Python engines. This should be done after the pyarrow engine is merged, and will help get rid of some code nested within an if statement. Ideally, we make a method on ParserBase, and call that in the C and Python engine read code so that read always returns a DataFrame now.

[phofl]

We should unify the parser-code base in general. A lot of code is duplicated in there

[jreback]

+1 all around here. good news is we have a gazillion tests so refactoring PRs should be straightforwrrd to review / merge.