ETL's Hivesql tuning (union ALL)

Believe that in the process of ETL inevitable practical union all to assemble the data, then this involves whether the problem of parallel processing.

Whether a parallel map is applicable in hive can be set by parameters:

set hive.exec.parallel=true

Then it is useful for the data of the previous blog, Link: http://www.cnblogs.com/liqiu/p/4873238.html

If we need some data:

Select  from (Selectfrom where create_time="2015-10-10"9718 Select  as  from where 9718) A;

Is the simulation from the two tables to obtain data, if not open parallel, practical time is twice times the opening time, so this place more attention!

ETL's Hivesql tuning (union ALL)