There are two main reasons for performing bucketing to a partition:
- A map side join requires the data belonging to a unique join key to be present in the same partition.
- It allows us to decrease the query time. Also, makes the sampling process more efficient.
Hope this helps