04-08-2014 04:05 AM
I'm trying to run the following code:
proc sql; connect to hadoop (user="myusr1" pw="mypwd1" server=hxpduped port=10000 schema=default); create table test as (select * from connection to hadoop (select * from dataset limit 100; )); disconnect from hadoop; quit;
But my code is running without any result. I can't able see the job running in job tracker as well.
Can anybody please help?
04-08-2014 05:01 AM
Have you completed the post install configuration, moving jar-files, setting up configuration file etc?
Have you tried a libname?
Can you ping the server?
04-08-2014 05:18 AM
Thanks a lot LinusH for your reply! Yes. We have completed all the configurations. I tried the libname option, but was getting the same problem.
We can ping the server as well. The Hadoop server is up & running. And we can submit task which are getting completed.
But can't able to connect sas with Hadoop. Also, I'm using SAS Enterprise in case that's an issue.
04-08-2014 07:45 AM
With libname, can you connect, but not read data? What does PROC DATASETS LIST give you?
Do you have a Hive or Hive2 server?
Have you tried FILENAME?
If you can connect but not read, it's...odd. And I guess that you tried the same HiveQL inside Hive with the same user?
Perhaps this typical SAS tech support issue, hard to troubleshoot remotely for forum participants...
04-08-2014 09:35 AM
Thanks LinusH for your helpful reply. We have Hive server. I also have the feeling that I need to contact SAS guyz to look into the issues. But thanks anywayz.
04-09-2014 11:36 AM
Can you try something similar to this and post your SAS log?
libname h1 hadoop server=hxpduped port=10000 user=myusr1 password=mypwd1 schema=default;
options sastrace=',,,d' sastraceloc=saslog nostsuffix;
select count(*) from h1.cars;