Background:
The segmentation application Audience Discovery and Targeting within the Customer Engagement Intelligence suite of HANA powered products is a
great tool for building segmentation models and target groups to feed downstream campaign execution processes.
A solution brief can be found here.
In a regular implementation project one would replicate the required data from connected SAP (e.g. ERP/ CRM/ BW...) or non-SAP systems
using i.e. SAP SLT or SAP BO Data Services. However I regularily get requests to set-up Proof of Concepts or Demo scenarios based on
customer specific data without having already the complete system landscape set-up that would allow loading data to SAP HANA via the
mentioned tools.
In this case we can use a CSV File and import the same as foundation for a customer specific segmentation model.
Detailed instructions:
Prerequisites:
We will import data into the HANA DB using an import wizard functionality that’ll create the new table in the HANA DB and fill the table
accordingly. As the system tries to derive the table element definitions (data types of fields) from the source file, it’s advisable to use a properly defined
and filled CSV File as starting point.
accordingly. As the system tries to derive the table element definitions (data types of fields) from the source file, it’s advisable to use a properly defined
and filled CSV File as starting point.
Tip:
- Use capitalized letters for your column names
- Do not use blanks in your column names
- Do not use German “Umlaute” such as ä, ö, ü in your column names
- Prevent from leaving cells completely empty. If somebody has no Hobby use “na” or “unknown”. If a numeric value is zero, fill the cell with “0”.
Following this might help to prevent issues during data upload and later during modeling and further data manipulation via SQL if required.
Upload the prepared data files to your HANA DB with SAP HANA Studio
We assume you already have a defined HANA DB schema that will contain the new table after upload. In our case, we use REINERM.
Upload wizard is launched from the quick access view. Select -> SAP HANA Content -> Data from local file -> Import.
Select File from your local machine. The Field Delimiter for our CSV File is “Semi Colon”.
Our CSV File contains a first row with column names -> “Yes”, Header row exists.
We create a new table in the existing schema. If the table was created already e.g. manually in HANA via SQL
statement, the “Existing” option would apply. -> Next.
statement, the “Existing” option would apply. -> Next.
The system proposes a field mapping and data types for the fields of the table to be created. Check this, adjustments might be advisable. The proposed data types might be inappropriate i.e. “City” is proposed as “NVARCHAR14” but in your CSV File row 999998 contains an entry 20 characters long. If you do not manually adjust the data type you’d get in import error.
PARTNER_ID -> change to NVARCHAR10, in the master table each business partner has only one record, therefore we can define the PARTNER_ID as Key field for the table.
You can also use the preview to check if based on the field definitions the data is displayed correcly. If yes -> next -> finish.
Check Job Log -> should say “Completed successfully”.
Result: Table should be created with data uploaded. Check Data Preview.
Repeat the same steps for the dependent table (!!! In this Case GUID is the key field as we have potentially more than one record per PARTNER_ID!!!).
Now view modelling for a custom ADT Segmentation model can start.
HANA view modeling
HANA Views are organized in the Content Area -> Packages. It’s advisable to create a new package.
Tip: Do not use underscore (_) for the package name, it causes issues later in CEI customizing when defining the data source.
Now we can create the first Attribute view (right mouse click the package name).
Add the required tables to the Data Foundation area of the attribute view (drag&drop).
Set-up relation between key fields (drag&drop) – for these two tables PARTNER_ID makes sense.
Adjust join properties -> Join Type: Left outer (view will return all BPs, even if they do not have dependent records).
Fields required as attibutes for segmentation have to be added to output structure of the view -> right mouse click -> Add to Output.
Adjust output field properties -> click Semantics
Every view requires a key attribute -> in this case PARTNER_ID as key attribute makes sense.
You can also adjust field labels. Just enter the required names in the "Label" column.
Save and activate view.
Data preview of the view should now return values. -> View is ready to use as data source in ADT customizing.
CEI-ADT Customizing
Enter application the implementation guide in the CEI application server (wellknown transaction SPRO).
Complete customizing steps in Node Segmentation. The IMG documentation is pretty clear. An overview of the required steps can be seen below.
Steps 1.1 and 1.2 are done with the HANA related tasks described above. Steps 2.1ff are all application server/ IMG related.
Have fun with your custom CEI-ADT segmentation scenario!