[Bug]: DataFrame API `to_feather` fails in most cases

### What happened?

Pandas to_feather fails when used with a non-default index:
```
ValueError: feather does not support serializing a non-default index for the index; you can .reset_index() to make the index into column(s)
```

Unfortunately, our DataFrame partitions will only have a default index in rare cases. Typically each partition will have some hash-based index. So to_feather will usually fail with the above error at execution time.

- https://stackoverflow.com/questions/74345723/how-to-save-deferreddataframe-to-feather-with-dataflowrunner
- https://lists.apache.org/thread/wszft7ts7l4xg05lqfbwybn1mbpkkffz

### Issue Priority

Priority: 2

### Issue Component

Component: dsl-dataframe