Data Modeling - Pivot
1 - About
The principal article on the pivot (format|analysis) or structure.
2 - Articles Related
3 - Disadvantage as a storage structure
Disadvantage of this structure as a storage structure, if a pivot is used to represent different possible values of a column.
If you want:
- to add a category, you need to add a category column for each column
- to do an update on one of this fields, you end up with a list of “IF THEN ELSE”.
- to do an aggregation on one of this fields, you end up with a list of “UNION”.
- to add an audit function on it, you need to add an audit on each column
- to set a value that is dependent of others dimensions, you end up adding a lot of column.
- If you have 1 metrics column with 2 dimensions, you add 2 columns
- If you have 2 metrics column with 2 dimensions, you add 4 columns
Furthermore, it's not a third normal form. Then if the data is sparse, you cannot save it in a parse form and you end ups using a lot of storage for nothing and can degrade seriously then the performance.