-
1. Re: Getting error while pulling columns after connecting to Hortonworks Hadoop Hive
Prashant Sharma
Apr 17, 2015 3:34 AM
Hi Sindhu,
Check this thread - Tableau 9 Desktop and Hive
Also, I think your ODBC driver is not supporting when HIVE definition when creates a logical object in conjunction with the use of a custom JSON-SerDe.
Warm Regards,
-
2. Re: Getting error while pulling columns after connecting to Hortonworks Hadoop Hive
Hi Prashant,
Thank you so much for your quick response.
I tried to use hdfs username and I was able to connect to the datasource, but I am getting error when I am trying to pull the column in the report.
Please let me know for any other inputs.
Thanks
Sindhu
-
3. Re: Getting error while pulling columns after connecting to Hortonworks Hadoop Hive
Prashant Sharma
Apr 23, 2015 1:18 AM
Not sure but I think that ODBC is not supporting logical objects. Let me ping Russell Christopher. He may suggest something in this.
Warm Regards,
-
4. Re: Getting error while pulling columns after connecting to Hortonworks Hadoop Hive
Has anyone found a resolution to this? I am facing the same issue when using a Live connection to Hortonworks Hive. Extracts work ok.
-
5. Re: Getting error while pulling columns after connecting to Hortonworks Hadoop Hive
Hello Mark!
Idon't have a solution for this issue, but this Does not occur in tableau 9.3.
I'm using:
Windows 7 (64 bits)
Tableau 9.3
Driver ODBC (HortonworksHiveODBC64) you can download in this link http://public-repo-1.hortonworks.com/HDP/hive-odbc/1.4.14.1041/windows/HortonworksHiveODBC64.msi
Regards
-
6. Re: Getting error while pulling columns after connecting to Hortonworks Hadoop Hive
Sindhu Srinivas Not sure if you already found a resolution or not, but here is this thing which worked for me.
To recap,I noticed that I was using my login credentials as "A123****" and Tableau was establishing the connection and showing the data preview but threw 'ExecuteStatement finished with operation state: ERROR_STATE' error upon pulling the data.
I changed my login credentials to lowercase "a123***" and to my delight there was no ERROR_STATE error any more.
Upon further investigation I learned that this problem was related to Hadoop's Map Reduce Job execution requirement. To put it simply, the MR jobs needs a user directory (with exact case sensitive name) to store temp files which are generated as part of the execution process. So when we just preview the data, no such job kicks in and HIVE does not look for temp storage in the user directory but when we pull the data for dash-boarding / reporting purpose that is when HIVE actively seeks a exact math user name and throws the ERROR_STATE error if it fails to find one.
Hope this helps rest of the user base.