ERPJEWELS.COM

Sr

Relational Modelling

Multi-Dimensional Modelling

1

Relational models can be very complex with hundreds of tables having long chains of relationship among them.

Multidimensional modeling are very simple. Each of the dimension table has a direct relationship with the fact table

2

Normal data modeling is quite flexible. 

The Multidimensional modeling has a rigid structure

3

One of the goals of relational modeling is to confirm to the rules of normalization. In a normalized database each data value is stored only once.

Multidimensional modeling are radically denormalized. The dimension tables have a high number of repeated values in their fields.

4

Standard relational models are optimized for On Line Transaction Processing. OLTP needs the ability to efficiently update data. This is provided in a normalized database that has each value stored only once.

Multidimensional modeling are optimized for On Line Analytical Processing. OLAP needs the ability to retrieve data efficiently. Efficient data retrieval requires a minimum number of joins. This is provided with the simple structure of relationship in a Multidimensional modeling, where each dimension table is only a single join away from the fact table.

5

SQL queries required to manipulate relational data

MDX queries required to manipulate multi-dimensional data

6

Tables are units of relational data storage.

Cubes are units of multi-dimensional data storage.

7

Table fields of particular data type store the actual data.

Dimensions and measures stores actual data.

8

Table size is measured in number of record set.

Cube size is measured in number of Cellsets. In this sense, MS Excel Pivot Table can accommodate 2,040,000 cellsets (8000 * 255). Considering the summary data to be reported, this is reasonably large. Pivot Table can report at the most 8000 different items.