Data Block & Index System

Data Block (Building Block of Essbase) & Index System

Data values are defined as the intersection of one member from one dimension with one member from another dimension. A data value is stored in one cell in the database.  In order to refer to that specific data value in a multidimensional database you need to specify its members from each dimension.

  

Essbase stores and accesses data through the use of data blocks and the index system. A data block is created for each unique combination of sparse dimension members.  Essentially, the data block represents all the dense dimension members for its combination of sparse dimension members.

Every time a data block is created, Essbase then creates an index entry.  The index is comprised of combinations of sparse dimension members.  There is one entry for each combination of sparse dimension members where a data value exists. The data block is a fixed format data structure the existence of which is driven by data-relevant sparse member combinations in the index. By data-relevant we mean that only where business data actually exists across sparse member combinations will a data block be generated.

In the example below, Entity, Scenario, and Year are sparse dimensions, while Accounts and Periods are dense dimensions.  When Essbase searches for a data value, it is using the index to locate the block containing that data value.  Once the data block is located, it targets the exact cell containing the data value.  Essbase is able to handle sparse data so well because the index provides a pointer to the correct data block.  Once the data block is located, Essbase can quickly retrieve the data value.  In the example below, let’s say you are searching for Budgeted Expense for NY in February 2013.   The index provides a pointer using the sparse dimensions to locate the data block containing the data for NY, Budget, 2013. The data value in question is then contained at the intersection of Expense and Feb.