Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Hive >> mail # user >> ENV variables from CLI


Copy link to this message
-
Re: ENV variables from CLI
Thanks guys,    Adding the 'env:'  in my 'add jar'  works.  

Sam
On May 3, 2012, at 7:35 AM, Edward Capriolo wrote:

> That is generally how you set hiveconf. Env variables can be accessed this way.
>
> hive> set x=${env:HOME};
> hive> set x;
> x=/home/edward
>
>
>
> On Thu, May 3, 2012 at 2:48 AM, Nitin Pawar <[EMAIL PROTECTED]> wrote:
>> if you are looking at having parameters from external world to hive shells
>> then i would recommend
>>
>> 1) define all your query set into a file
>> 2) when you execute the query like hive -f hive.hql  -hiveconf
>> parameter1=`echo $env-value`
>> and then in the hive.hql file you can access this command line parameter
>> with '${hiveconf:parameter1}'
>>
>> ~Nitin
>>
>> On Thu, May 3, 2012 at 11:53 AM, Sam William <[EMAIL PROTECTED]> wrote:
>>>
>>> How can I read env variables from the CLI .  The basic set command might
>>> not be enough for me.
>>>
>>>
>>> Essentially this is what Im trying to do - create a .hiverc file   under
>>> $HIVE_HOME/bin  so as to to create UDFs for everyone in the company to
>>> access .   I d like to create the .hiverc file with  commands like
>>>
>>>  add jar  ${JAR_LOC}/custom_hive_udfs.jar;
>>>  CREATE TEMPORARY function  BlastIt as 'com.test.hive.udf.BlastIt';
>>>
>>> Id prefer not to  hardcode the location of  the jar file.   What are my
>>> options ?  Generating the .hiverc file everytime there s a change  can only
>>> a second-best option .
>>>
>>>
>>> Sam William
>>> [EMAIL PROTECTED]
>>>
>>>
>>>
>>
>>
>>
>> --
>> Nitin Pawar
>>

Sam William
[EMAIL PROTECTED]