Index name is not respected properly in staged writes

[G-D-Petrov]

          The field count will also be `0` for range indexes. I just checked and range indexes can have a name in Pandas, but I don't know if we persist them, can you check?

Originally posted by @alexowens90 in #2116 (comment)

[G-D-Petrov]

@pytest.mark.parametrize("append", (True, False))
@pytest.mark.parametrize("delete_staged_data_on_failure", [True, False])
def test_parallel_dynamic_schema_named_range_index(
    lmdb_version_store_tiny_segment_dynamic, append, delete_staged_data_on_failure
):
    lib = lmdb_version_store_tiny_segment_dynamic
    sym = "test_parallel_dynamic_schema_named_range_index"
    df_0 = pd.DataFrame(
        {"col_0": [0], "col_1": [0.5]}, index=pd.RangeIndex(1)
    )
    df_0.index.name = "date"
    df_1 = pd.DataFrame({"col_0": [1]}, index=pd.RangeIndex(1))
    df_1.index.name = "index"
    if append:
        lib.write(sym, df_0)
        lib.append(sym, df_1, incomplete=True)
    else:
        lib.write(sym, df_0, parallel=True)
        lib.write(sym, df_1, parallel=True)

    with pytest.raises(SchemaException) as exception_info:
        lib.compact_incomplete(
            sym,
            append,
            False,
            delete_staged_data_on_failure=delete_staged_data_on_failure,
        )

    assert "date" in str(exception_info.value)
    staged_keys = 1 if append else 2
    expected_key_count = 0 if delete_staged_data_on_failure else staged_keys
    assert len(get_append_keys(lib, sym)) == expected_key_count

This test should pass because the 2 df that need to be compacted have different names.
Currently it doesn't fail, because the index names for range indexes are not passed through to the cpp layer and only the name of the first one is stored in the normalization metadata.