limit and distinct don't work together properly

### Description

It seems that:

```python
for i in range(loops):
        c = dc.read_dataset(dataset_name).limit(limit).distinct("session_id", "person.position").count()
        print(f"loop_{i}:", c)
```

return same `limit`, even if there are duplicates. persist() after `limit` helps.



### Version Info

```Text

```