Hi Mahsa,
Just to elaborate a little further.
Bucket joins are quicker than regular joins because they lessen the number
of logical multiplications that need to happen between the records from two
tables being joined.
A regular bucket join would logically multiply each row from one table with
each
The number of mappers is the greater common factor for all the tables.
So, it it is not a multiple, a single mapper will be assigned.
On 11/1/12 5:20 AM, "Mahsa Mofidpoor" wrote:
>Hi,
>
>Hive summit 2011 says for performing a bucket join the number of buckets
>of
>all tables has to be a multipl
