Thanks Kunal for clarifying... I am still learning the things, so as per my first project I am trying to create a successful connection. I will work on the optimization in the second phase. Thanks for your valuable tips, let me try to create the hive connection through JDBC then. I suppose I need to put the hive jdbc drivers in the 3rd party directory, please let me know if you have the list of the drivers or jar I need to put in the 3rd party directory.
Thanks Asim On Wed, Mar 7, 2018 at 6:06 PM, Kunal Khatua <kunalkha...@gmail.com> wrote: > You should be able to connect to a Hive cluster via JDBC. However, the > benefit of using Drill co-located on the same cluster is that Drill can > directly access the data based on locality information from Hive and > process across the distributed FS cluster. > > With JDBC, any filters you have, will (most probably) not be pushed down > to Hive. So, you'll end up loading the unfiltered data through a single > data channel from the Hive cluster, into your Drill cluster, before it can > start processing. > > If using JDBC is the only option, it might be worth using the 'create > table as' (or the temporary table variant) to offload that data into your > Drill cluster and then execute your analytical queries against this > offloaded dataset. > > On 3/7/2018 2:46:55 PM, Asim Kanungo <asim...@gmail.com> wrote: > Hi Team, > > Can I connect to hive database as a generic jdbc protocol like I am doing > for other RDBMS.Or DRILL can only connect to hive residing in the same > cluster where Hadoop is installed. > > I am talking about the cases where DRILL is installed in one cluster and my > Hadoop cluster is different. Can you please guide if I can connect to that > hive, using jdbc protocol. > I have tried, but it failed with the error "unable to add/update storage". > I am not sure why it has failed as it is not giving any other messages, and > I know version 13 will contain more message. > > Please advise. > > Thanks > Asim >