[C++][Dataset] Ensure that dataset code is robust to schemas with duplicate field names

[asfimport]

Fields in a schema don't have to have unique names, so we should make sure we're not assuming that.

Reporter: Neal Richardson / @nealrichardson
Assignee: Ben Kietzman / @bkietz

Related issues:

• [C++][Dataset] Handling of duplicate columns in Dataset factory and scanning (relates to)

PRs and other links:

• GitHub Pull Request #6545

_{Note: This issue was originally created as ARROW-7412. Please see the migration documentation for further details.}

[asfimport]

Ben Kietzman / @bkietz:
For now, we should probably just error if a field is referenced whose name is not unique.

[asfimport]

Ben Kietzman / @bkietz:
A more general solution would be provide an unambiguous way to refer to fields https://docs.google.com/document/d/1-aoAb1d00BfCcEP--kmYWDBb_eeOFgtlSVPPjSGy8ko/edit

[asfimport]

Ben Kietzman / @bkietz:
Issue resolved by pull request 6545
#6545