Create IDT universe from Apache Hive data source

building semantic layer
1

Hi there

I am trying to create a universe (using IDT) from an Apache Hive data source. I have managed to create both a JDBC connection and an ODBC connection to the data however neither works :cry:

The JDBC connection allows me to see the schema, but no tables within the schema and the ODBC connection allows me to see internal tables and include them in the Business Layer, but I cannot access any of the data in the table - view/profile etc. I also cannot see any external tables or views.

This is a POC that we’re working on and it would be great to prove that BO can access this data correctly.

The BO system is BO 4.1 SP5
The JDBC connection was created using Apache Hadoop Hive 0.13 Simba JDBC4 HiveServer2
The ODBC connection was created using Apache Hadoop Hive 0.13 Simba ODBC

I haven’t loaded any jar files or anything as I was led to believe that the newly released drivers with 4.1 sp5 don’t require it. Most of the documentation I can see is from 2013/2014 and I’m not sure it’s still valid. The data is sitting on a VM server on a Linux machine.

Any assistance would be gratefully received

SarahG :australia: (BOB member since 2005-08-22)

2

First of all…if is a PoC DEFINATELY get with a recent version…4.1 SP5 is really kinda old.

Go with 4.2 SP2 or 4.2 SP3.
then get a recent IMPALA driver or VORA… I would guess…

And have you seen this post: https://wiki.scn.sap.com/wiki/display/BOBJ/Apache+Hadoop+Hive+Connection+With+Information+Design+Tool

Andreas :de: (BOB member since 2002-06-20)

3

Hi there
Thanks but they are really sticky about what versions I can use and they don’t want the hassle of an upgrade at this stage - there are over 5000 reports and they would insist on regression testing all of them!!!
I’ll look into it in a new server maybe but my hopes are not high.
Can I use an Impala driver on a MapR solution? Does it matter what the underlying infrastructure is??

SarahG :australia: (BOB member since 2005-08-22)

4

Oh well…wanna use new functions & features but not get on a recent version, and rather stay with an outdated one, which is now more than 2 years old ? Oh well… :shock:

Anyway, have you seen this SAP Note: https://launchpad.support.sap.com/#/notes/0002051331

Andreas :de: (BOB member since 2002-06-20)

5

Yep - they keep saying they might not be using BO going forward so they don’t want to spend too much time upgrading it. The end users basically use it as an extract tool to get the data into Excel and manipulate it there! There are lots of other tools being bandied about but nothing that has a semantic layer like BO. I’m new here and trying to change their minds on a few things :nopity:
I’m going to suggest we upgrade an unused dev server and access the Hive data from there.
I can’t access the link you sent as the S-ID user password has expired and we’re waiting to get it updated! Things don’t always work as quickly as we’d like :blue:

SarahG :australia: (BOB member since 2005-08-22)

6

Still, all the best… I’ll keep my fingers crossed for you :+1:

Andreas :de: (BOB member since 2002-06-20)

7

SarahG, don’t know if you ever got this working, but the Data Access Guide (for BI 4.2 Sp04) says you need to copy a set of JAR files to \connectionServer\jdbc\drivers<hive012simba4server1>.

charlie :us: (BOB member since 2002-08-20)