Updating olap cubes sql server pains wessex website
Especially if you want to use only some parts of it for any kind of calculation.When can you store serialized data: When it is not important to enforce the consistency and you will NEVER use it for statistical queries or for filtering.Referencing and updating columns is better that updating and referencing JSON fields.In this case you are optimizing DML performance and probably some analytic.Be sure you're doing what's best for your application, and not simply avoiding the object-relational problem because you want to.More attention needs to focus on what @Samcd questioned.
One of our team members is suggesting that we store our row data into a JSON column rather than in multiple SQL columns.We will have to report off of these line items so performance is definitely a concern.We also are running SQL Server 2012 which as I understand does not contain native support for JSON columns unless we upgrade to SQL Server 2016.Storing multiple data in one field also means, you can not (or not easy to) As an alternative, you can use XML columns, it is supported by SQL Server and some of the above things are not a problem, BUT: it still has no fixed structure.(if it has, you can store the data in traditional columns.
The cubes immediately incorporate changes that are made to the data in the database, based upon notifications received from the database.