Hi all,

Hive already have query (statement) level pre/post execute hook, I think
implementing job level pre/post execute hook is also needed, based on the
reasons listed below:

1. The `set key=value' statement changes session level configurations, not
able to change job-wise config like "hive.groupby.skewindata". If a single
query needs 2 different config for its different jobs, we have to insert
temporary tables first.

2. Its better to introduce some post execute analyze tools like "vaidya"
when job level hook is implemented.

Any ideas?

Best Regards,
Ted Xu

Search Discussions

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupuser @
categorieshive, hadoop
postedJan 4, '11 at 3:05a
activeJan 4, '11 at 3:05a

1 user in discussion

Ted Xu: 1 post



site design / logo © 2022 Grokbase