Hello Tableau Enthusiasts and Experts
Scenario: My organization is trying to migrate all reports and dashboards from Microstrategy to Tableau, and do away with Microstrategy. I know both are different tools one is reporting and other is visualization, but we are trying to explore what best can we achieve. We have multiple databases each containing hundreds of tables. Our user base is habitual of using Microstrategy metadata server for pulling attributes and metrics for their ad-hoc reporting needs and they are very comfortable using the metadata as they do not have to care about what is going on behind the scene. They do not need to know the tables to choose for their need nor do they need to know how the tables need to be joined to each other.
We are trying to build something similar in Tableau. What we could gauge from different forums and whitepapers is that the "Tableau Datasource" suffices as the metadata layer for users to start building their workbooks on top of published data sources.
Problem Statement: My question is, is it possible to create a single data source joining 100's of tables from multiple databases, in order to have that as the central repository for all the users to use. If my user wants to look at sales by state, will Tableau behind the scene queries only two tables (sales from fact and state from territory dimension table) or will it query and join all the 100 tables involved in the data source. Will it not deteriorate performance drastically? What is the best way to handle this situation?
Hi,
You're trying to use Tableau with a Microstrategy mental model, I don't know Microstrategy however my experience with migrations from various other tools has been that trying to use the mental models of the old tools without learning how Tableau works (and therefore Tableau's mental models) leads to lower quality outcomes in terms of user satisfaction, time to implementation, performance, etc.
Tableau simply doesn't have the functionality that you are describing where the metadata is such that Microstrategy will join anything to anything and (presumably) that "just works" for users for aggregations, performance, etc. Tableau is continually evolving the product so it may have more of that functionality in the future (pinging Roger Hau) but that's not where Tableau is at today.
What I've seen companies do on the data side is to take some time to understand the use cases and user requirements and build multiple data marts (and potentially published Tableau data sources) where each data mart or published data source handles one or more use cases.
Jonathan
