arrow

ARROW-10732: [Rust] [DataFusion] Integrate DFSchema as a step towards supporting qualified column names

This PR builds on https://github.com/apache/arrow/pull/8840 and integrates DFSchema with the DataFusion query planning, optimization, and execution code.

This was a pretty large refactor unfortunately and I don't really see a way to break this down into smaller PRs.

There should be no functional changes in this PR. Fields are looked up using `field_with_unqualified_name` and I will file a separate PR to add support for referencing qualified field names.

Note that I had to update `PhysicalExpr.evaluate()` to pass in the input schema since we can no longer rely on the schema from the Arrow `RecordBatch` (because it loses the qualifiers). The other methods on `PhysicalExpr` already required the input schema, so this seems consistent at least, because we now always use the schema from the plan.

The rest of the changes are updating the query planning, optimization, and execution to use `DFSchema` instead of `Schema`.

Design Document: https://docs.google.com/document/d/1BFo7ruJayCulAHLa9-noaHXbgcaAH_4LuOJFGJnDHkc/edit#heading=h.3japu7255aut

Closes #8839 from andygrove/sql-relation-names

Authored-by: Andy Grove <andygrove73@gmail.com>
Signed-off-by: Andrew Lamb <andrew@nerdnetworks.org>