Beeline Configuration hiveconf Equivalent

[#1]

Need to set hive configuration options from Beeline CLI command. The below code, for example, doesn't work.

--hiveconf hive.execution.engine=tez

--hiveconf and --property-file both don't appear to be making the session take the settings.

Here it is in full usage:

beeline -n $user -p $password -u "$jdbc_url" -f $script --verbose true --property-file "query.properties" --fastConnect

Is the only way to make it work to include the options in the JDBC URL as arguments? Like hive.execution.engine=tez;next;next ?

[#2]

you can use set hive.execution.engine=tez; while in a session or beeline --hiveconf hive.execution.engine=tez should work. what version of hive are you using? Are you setting any other value for --hiveconf?

Another option is the sel set hive.execution.engine=tez; inside your hive script or have --hivevar has parameter into your hive script to set execution engine

beeline -u jdbc:hive2://hostname:10000 -n xxxx -p xxxx -f /home/hdfs/scripts/hive/myscript.hql --hivevar engine=tez

Then use ${engine} variable inside your script.

[#3]

Using a set works fine, correct. However, that requires that I do that for every script I try to automate. My goal is to make these values a default part of this base command, and just change what SQL I throw at it. Global values, if you will.

This should be remedied by the use of --hiveconf or --property-file, but neither results in it taking the values. Here's an example:

Command:

beeline -n $user -p $password -u "$jdbc_url" -f $script --verbose true --property-file "query.properties" --fastConnect

query.properties

hive.execution.engine=tez;
tez.queue.name=di;
hive.exec.parallel=true;

The job shows up in the 'default' queue when it should show up in the 'di' queue. When I use a set command, it does in fact show up in the 'di' queue.

[#4]

what version of hive are you using?

[#5]

Hive-Hcatalog 1.2.1

[#6]

I just tried --hiveconf and it did not set the execution engine. Digging why it didn't take.

[#7]

I got it working now. Here is what I did

[LAKE] [smanjee@lake1 ~]# beeline -u "jdbc:hive2://example.com:2181,example.com:2181,example:2181/;serviceDiscoveryMode=zooKeeper;zooKeeperNamespace=hiveserver2" --hiveconf hive.execution.engine=mr

and it worked. I was missing the " " around the jdbc

0: jdbc:hive2://example.com:2> set hive.execution.engine;
+---------------------------+--+
|            set            |
+---------------------------+--+
| hive.execution.engine=mr  |
+---------------------------+--+
1 row selected (0.192 seconds)
0: jdbc:hive2://example.com:2>

[#8]

WORKING WORKAROUND:

Set the options in your JDBC URL. So your URL goes from this...

Old:

jdbc:hive2://hadoop.company.com:8443/default;transportMode=http;httpPath=gateway/default/hive

New:

jdbc:hive2://hadoop.company.com:8443/default;transportMode=http;httpPath=gateway/default/hive?hive.execution.engine=tez;tez.queue.name=di;hive.exec.parallel=true;hive.vectorized.execution.enabled=true;hive.vectorized.execution.reduce.enabled