Projection pushdown produces incorrect results when column names are reused

[jonmmease]

Describe the bug
In certain situations where projections are used to overwrite a column's value, the projection pushdown optimization pass causes the incorrect column value to be produced.

To Reproduce
Here is a failing test that will be included in the follow-on PR

#[tokio::test]
async fn select_with_alias_overwrite() -> Result<()> {
    let schema = Schema::new(vec![Field::new("a", DataType::Int32, false)]);

    let batch = RecordBatch::try_new(
        Arc::new(schema.clone()),
        vec![Arc::new(Int32Array::from_slice(&[1, 10, 10, 100]))],
    )
    .unwrap();

    let ctx = SessionContext::new();
    let provider = MemTable::try_new(Arc::new(schema), vec![vec![batch]]).unwrap();
    ctx.register_table("t", Arc::new(provider)).unwrap();

    let df = ctx
        .table("t")
        .unwrap()
        .select(vec![col("a").alias("a")])
        .unwrap()
        .select(vec![(col("a").eq(lit(10))).alias("a")])
        .unwrap()
        .select(vec![col("a")])
        .unwrap();

    let results = df.collect().await.unwrap();

    #[rustfmt::skip]
        let expected = vec![
        "+-------+",
        "| a     |",
        "+-------+",
        "| false |",
        "| true  |",
        "| true  |",
        "| false |",
        "+-------+",
    ];
    assert_batches_eq!(expected, &results);

    Ok(())
}

presently, the values produced for column a are the original values [1, 10, 10, 100] rather than the expected values [false, true, true, false].

PR to follow shortly