Hadoop Hands-on Part 2 Impala Testing

We were having difficulty getting a SQL query copied from our EDW to work correctly in HUE using Impala under the Query Editors Menu.


The query is ugly and one of our worse performing so it is a great candidate for Hadoop and specifically Impala. I tried the query in HUE and it would begin to execute and reach 81% completion and then it would stop and show the results tab with a record count but showed a status of executing in Cloudera Manager.

I decided to remove the GUI as the possible issue and used Impala-Shell which is a common troubleshooting technique. I pulled out the Impala Manual, logged into the shell, did some quick testing, and then executed the query as a script and successfully loaded 86 Million records in 3 minutes.  I have a WebEx setup to troubleshoot why this does not execute successfully in HUE but the results are still impressive.

Leave a Reply

Your email address will not be published. Required fields are marked *