CDS Views vs. Extractors vs. SAP Tables
Ingesting via Extractors/CDSViews
- Utilizes existing SAP-provided business logic.
- Incremental/delta data loading might be challenging, especially for larger data sources.
- Not all CDSViews support extraction or incremental data loading.
- Certain older SAP Extractors may not sustain frequent data pulls (exceeding one per hour), which may have a serious performance impact on the source S4 system.
- Most CDSViews do not contain much logic on the top of RAW SAP tables
- Extra level(s) of abstraction which make things more complicated
Ingesting RAW SAP ERP tables
- Predictable performance for large SAP data sources
- Incremental data loading works without compromises.
- No/Low Compliance issues for the future
- SAP-provided business logic must be fully re-engineering in Databricks.
- Requires more experienced SAP experts
- SAP SLT license might be needed.
Example Extraction Annotations
I_GLAccountLineItemRawData CDS View built on the top of the ACDOCA table
→ Read more on how to set up incremental load for CDS Views on SAP website.
@Analytics: {
dataCategory: #CUBE,
internalName: #LOCAL,
dataExtraction: {
enabled: true,
delta.changeDataCapture: {
mapping:
[
{
table: 'ACDOCA',
role: #MAIN,
viewElement: ['SourceLedger', 'CompanyCode', 'FiscalYear', 'AccountingDocument', 'LedgerGLLineItem'],
tableElement: ['rldnr', 'rbukrs', 'gjahr', 'belnr', 'docln']
}
]
}
}
}