You can translate the question and the replies:

Aggregate Awareness

Related to: Data Virtualization and ETL
Not sure if this question belongs in this section, but I would like to know how Denodo handles aggregate awareness. Does Denodo have the ability to be configured logically to pull from a single fact table, but physically have multiple tables at different grains - and know the best table to pull data from. For example, if I have tables fact_detail, fact_hour and fact_day the difference between the tables are that the detail table has a timestamp, the hour table has a date field and an hour field, and the day table only has a date field. If I write a query that sums by day, I want it to only query the day table and if I add the hour attribute it should query the hour table, etc. Thanks.
Aggregate query re-write
user
03-04-2018 17:59:43 -0400
code

1 Answer

Hi, Take into account that reporting layer functionality isn't equivalent to what can be done in a relational model based on tables and views. I've been investigating about aggregate awareness and I've evaluated a couple of options depending on the scenario. On the one side, you could define the logic of the aggregation awareness using the partitioned union optimization, so Denodo Platform would go to the original table or to the aggregate one. I've tested this solution for simple scenarios and it worked as expected. On the other side, you could use custom policies. A custom policy could detect the restrictions of the aggregate table, and modify the VQL before the execution time. So, the current implementation of the custom policy could do the job although making the logic of those custom policies quite complex. Hope this helps!
Denodo Team
09-04-2018 05:09:23 -0400
code
You must sign in to add an answer. If you do not have an account, you can register here