Business Intelligence and NoSQL

Question

For a new project I am working, I need to store some data in a database. The data and access model type is well suited for NoSQL databases and databases. We look at things like MongoDB, but are still uncertain.

One of the requirements is the ability to use Business Intelligence tools to use some of our data. Others that I'm working on indicate that these tools allow you to use data directly from relational databases. I know that I don’t know which Business Intelligence tools I need to get.

Do BI tools typically support databases other than SQL? Are there other standard integration options that they have (some seem to support flat file web services) that might be applicable? Should I choose a relational database to make it easier to use with a third-party BI tool?

Answer 1

Do BI tools typically support databases other than SQL?

Not. And they really can't. Databases other than SQL are very different from each other. NoSQL is really just a collection of "things, not SQL."

However, the problem you are facing is classic. Most NoSQL databases run under the concept that you can “upload” and “output” the reports you plan to do. (usually this means reducing the map)

So, if you want to use common BI tools, you just need to "bridge the gap."

MongoDB supports CSV export, which can then be imported into a relational database for analysis. For some complex queries, you will most likely have to write “reduce map” queries and then export this data. Obviously, you have to write a couple of export / import scenarios.

But it’s entirely possible to provide BI reporting and use MongoDB as your primary data warehouse. You will still need RDBMS for the BI suite. But this is unlikely to require a lot of hardware, as it is simply used for reporting, not live data.

Answer 2

MongoDB is very nice, but it is not flexible. The developer determines the relationship between the objects when developing the application. However, the business intelligentsia means that people with domain knowledge can look for new relationships between entities. Data can get a new value that was not known when the developer designed the table / collection structure.