Hive Accumulo Integration

418 views Asked by At

Looking at

https://cwiki.apache.org/confluence/display/Hive/AccumuloIntegration

I'm wondering why there's no support for cell level visibility. Any thoughts?

1

There are 1 answers

0
C. Timms On

It's because Hive works with Accumulo by creating a Hive table based on an existing Accumulo table, allowing you to perform Hive queries on that data.

Unfortunately Accumulo's cell level security relies pretty heavily on the way Accumulo tables are structured and how scans are performed to work. Mapping it to a Hive table is just really impractical in a lot of ways. Instead Hive tables created from Accumulo data by performing a scan as an Accumulo user. Whatever data is visible to them will appear in the Hive table with no further security checks.

Ultimately if the cell level visibility feature of Accumulo is an important part of your application using Hive to make queries easier might not be the best idea.