HIVE login and show databases

Follow / 15 Feb 2012 at 9:22am

Hello All,
I installed hive server on one of the hadoop client boxes and i have toad running on windows. i was able to connect to hive server. However, i am not able to see any hive databases except default. Also, If i execute any query on default database, hive server throws following error
Authorization failed:No privilege 'Select' found for inputs { database:default, table:test..}. Use show grant to get more details.

If i login on client, i generally sudo to different accout and do hive queries. Is there a way i can achieve this and will be able to use different login than current login.

Replies

  • Follow / 16 Feb 2012 at 4:51am

    Hi,

    Our next release, due this month, will have support for connecting to non default databases on Hive.  We have also extended the Hive connection options to support SSH tunnelling which will allow you to connect as the target user with the desired privileges, which sounds like what your looking for.

    Regards,
    Stewart.

  • Follow / 20 Feb 2012 at 11:12am

    Thanks you very much for the info. In Upcoming release, will there be support for browsing and limiting different partitions?. Is there support for Joins? is there a way we can execute hive query itself without mapping a table?

  • Follow / 22 Feb 2012 at 8:09am
    Hi ravikumar,

    Unfortunately we don't have support for partitions in the upcoming 1.6 release of Toad for Cloud Databases (TCD). I would be interested in what specific information you would like to see, and if there are tasks you feel could be simplified by using TCD.

    In Data Hub connections we tend to stick to the formula of making everything look like a database that contains tables. This is so the user gets a consistent experience with mapped tables, regardless of the technology used by the remote data store.

    However, we have a specific interest in providing extended support for Hive with direct connections from the TCD client. These do not go through the Data Hub, and we would be more than happy to look at exposing partitions in these connections. Note that a direct connection does not require the Hive tables to be mapped before running a query.

    I may need more information in regard to your question on joins. The 1.5 product will execute HiveQL joins in both the Data Hub and direct connections. Could you give me an example of the sort of support you're after?

    Cheers,
    Greg.
  • Follow / 22 Feb 2012 at 10:15am

    Hello Greg,

    I am not able to get my queries working on hive.

    Our hadoop needs to set mapred.job.queue.name=validqueue for all queries to control th access. Is there a way we can set this propery in a hive query or data hub query.

    I am very much interested in querying with HIVE and joining with other databases like db2/mysql etc. Is that possible?

    First of all could you tell me set hive properties before the queries.

    i tried as below too and it doesnt work
    set mapred.job.queue.name=validqueue;
    select site_id, name from all_sites;

    It throws me error.

  • Follow / 23 Feb 2012 at 8:50am
    got it answered