hive-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Edson Ramiro <erlfi...@gmail.com>
Subject How Hive execute the stages?
Date Tue, 16 Jul 2013 18:00:00 GMT
Hi all,

I am executing the TPC-H [1] queries on Hive and I need help to understand
if Hive execute some stages locally. The TPC-H Query-16 [2] is translated
to three HiveQL queries, and the EXPLAIN [3] of each of these HiveQL
queries show me that the first query has 8 stages, the second query has 6
and the last has 4 stages. However, only 5 stages were submitted to Hadoop.

I think Hive does not submit some stages to Hadoop once these stages are
"internal" Hive operations like renaming tables, but I am not sure.

Would you please help me to understand what Hive does internally with the
stages? Does Hive execute some stages locally/at the master node? Why some
stages are not sent to Hadoop?

Thanks in advance,

       Edson Ramiro

[1] https://issues.apache.org/jira/browse/HIVE-600
[2] http://www.inf.ufpr.br/erlfilho/q16_parts_supplier_relationship.hive.txt
[3]
http://www.inf.ufpr.br/erlfilho/q16_parts_supplier_relationship.explain.txt

Mime
View raw message