did u do the performance recording and check where the bottle necks are?
How are you trying to connect to hadoop,i mean tableau and hadoop.
check the below link
"Tableau has developed a native connector to Impala, which makes it easy for any business user to quickly connect to their Hadoop cluster and begin doing analytics using Tableau’s visual drag-and-drop interface. Simply install the Cloudera ODBC driver and Tableau, connect to the cluster, and away you go! No Java or MapReduce programming required. Tableau takes care of generating the optimal queries needed to get the data out of Hadoop for the analysis and render it on the screen in the form of beautiful and rich data visulizations."
Hive is better than impala...
Impala is not 100% a substitute for Hive (Impala does not cover batch process and ETL, which are offered by Hive) but it is the option that offers shorter execution time in SQL queries as well as better integration with leader tools in
**** other links
Thank you for your reply.I have cloudera odbc driver installed in my local system.However I am connecting data source shared through tableau server.
I did run performance recording and found my queries were running in sequential not in parallel. I did go to the below link but i feel its for live connection.
Is there any way where I can enable parallel processing In my desktop and in server?
I suggest the following:
- I feel the performance fine tuning and parallel query processing has to be done primarily at impala server side
understanding if all the server clusters are properly utilized or not
if parallel query processing is enabled or not
- how your tableau workbooks are designed : a lot depends on that
what kinds of filters are used,Tableau order of precedence etc
- if u r using extracts or live connections it matters,live connections should be avoided
From tableau server finally the queries hit impala so it would be better if you check impala server first and then come to tableau.