Real-time fact tables need to be updated more frequently than the more traditional nightly batch process. There are many techniques for supporting this requirement, depending on the capabilities of the DBMS or OLAP cube used for final deployment to the BI reporting layer. For example, a “hot partition” can be defined on a fact table that is pinned in physical memory. Aggregations and indexes are deliberately not built on this partition. Other DBMSs or OLAP cubes may support deferred updating that allows existing queries to run to completion but then perform the updates.