Implementation
The extensibility implementation is a combination of functionality implemented in the LN-Model-Extensibility space and other LN-Model spaces. Before implementation, take these rules into consideration:
- You are the owner of the LN-Model-Extensibility space. This space will not be updated during an update of the LN Analytics when we start supporting updating the LN Analytics.
- In the Extensibility space, you are responsible to fill custom attributes for a specific dimension using Birst ETL. This also includes, for example, incrementally loading LN CDF fields from Data Lake. To automatically load the custom attributes into the data warehouse, the custom attributes must be filled into the Extensibility dimension in a fixed format.
This diagram shows an overview of the networked BI architecture for extensibility:
Note: For display purposes, some model spaces
are not included in the diagram.
The LN-Model-Extensibility space includes these packages:
- LN Extensibility, which contains the ‘Extensibility Dimension’ dimension, which is used by the extensibility logic. This package is imported in the LN-Model-ConformedDimension space and all functional area spaces.
- LN Extensibility- Labels, which is present for future functionality. This package and the related LN_csa_labels_customer source and CSA Labels Customer script should not be removed.
- LN Tables – Extensibility, which is imported from the LN-Extraction space. It contains, among others, the etl_table_sharing dimension table that includes the table sharing information of the LN tables available in the LN data sets.
The ‘Extensibility dimension’ contains these columns:
| Column | Data Type | Comment |
|---|---|---|
| Extensibility_Dimension_Key | Varchar(1100) | Used as Unique Identifier to make the record unique. Proposed
filling is: [Dimension] + ’*’ + [Level Key Column01] + ’*’ + [Level_Key_Column02] + ’*’ + [Level_Key_Column03] + ’*’ + [Level_Key_Column04] + ’*’ + [Level_Key_Column05] |
| Dimension | Varchar(50) | Column that should contain the exact Dimension name for which the custom attributes are added. |
| Level Key Column01 till Level Key Column05 | Varchar(200) | Columns that should contain the exact matching values of the Level Key column of the dimension. To be future proof five level key columns are added. |
| CDF Date01 till CDF Date05 | Date | The five date columns that you can fill for a dimension. |
| CDF DateTime01 till CDF DateTime05 | DateTime | The five datetime columns that you can fill for a dimension. |
| CDF Float01 till CDF Float05 | Float | The five float columns that you can fill for a dimension. |
| CDF Integer01 till CDF Integer05 | Integer | The five integer columns that you can fill for a dimension. |
| CDF Number01 till CDF Number05 | Number | The five number columns that you can fill for a dimension |
| CDF String01 till CDF String05 | Varchar(300) | The fifteen string columns that you can fill for a dimension. |
| sequencenumber | Integer | This column is used for incremental loading the extensibility data. If the dimension record already contains extensibility data, the extensibility data is updated only if the sequence number in the Extensibility Dimension is higher than the sequence number in the dimension. If there is no extensibility data present in the dimension, the extensibility data will also be processed. |
Note:
- You are responsible to fill the Extensibility Dimension with the correct data, so the data is picked up by the extensibility related scripts in the other LN-Model spaces. Therefore, it is important that the values in the Dimension and Level_Key_Column01 through Level_Key_Column05 are correct. For more details on the exact values of these columns, see Extensibility details.
- Once extensibility data is loaded for a dimension, the extensibility data can be updated only when the sequence number in the Extensibility column is increased for a particular dimension record.
- The LN-Model-Extensibility space is delivered with the ‘Extensibility Dimension’ script which can be used as a template to copy to other scripts. Consequently, you do not need to create all the columns and data types from scratch.