[Mondrian] Making Pre-aggregated Tables Work

Amit Shah amits.84 at gmail.com
Fri Apr 1 07:59:22 EDT 2016


Hello,

I have been using Mondrian 4 with Saiku as an OLAP engine. Mondrian is
integrated with apache phoenix (version 4.6) as the data source that used
apache hbase as the data store. In order to speed up the queries I am
trying to make pre-aggregated tables work but for some reason the actual
fact table is getting queried. Here's how the mondrian schema definition
looks like - http://pastebin.com/xquutKYX.

The database schema for the dimension, facts and the pre-aggregated table
is shared here - http://pastebin.com/EWhyNLTJ

Selecting the total sales measure and the agent id as a row in the saiku
builds up a query and tries to execute it on the facts table instead of the
pre-aggregated table. The query from the sql logs is

DEBUG [mondrian.sql] 850: Segment.load: executing sql [select
"AGENTS"."A_ID" as "c0", sum("TRANSACTIONS"."TOTAL_SALES") as "m0" from
"TRANSACTIONS" as "TRANSACTIONS", "AGENTS" as "AGENTS" where
"TRANSACTIONS"."AGENT_ID" = "AGENTS"."A_ID" group by "AGENTS"."A_ID"]

What could be missing in order to make the OLAP engine to use the
AG_TER_PRO (pre-aggregated table)?

Thanks,
Amit.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.pentaho.org/pipermail/mondrian/attachments/20160401/01cfd032/attachment.html 


More information about the Mondrian mailing list