Above-average programming depth for a BI architect — ABAP, SQL and AMDP used deliberately, each where it is strongest.
ABAP runs on the application server and excels at orchestration, interfaces and business logic. SQL is set-based and lives in the database. AMDP — ABAP-Managed Database Procedures — bridges the two: SQLScript written and versioned inside ABAP classes, executed natively in HANA.
Classic ABAP pulls data to the code: loops over internal tables, row by row. The HANA paradigm pushes code to the data: joins, aggregations and calculations run in parallel inside the database, eliminating costly round trips.
Process control, authorizations and interfaces stay in ABAP. Straightforward reads belong in SQL and CDS views. Complex, mass-data transformations — in BW transformation routines or planning functions — are AMDP territory.
I reimplemented a grown ABAP / BI-IP solution for tax-driven cost allocations as a fully HANA-based model with complex AMDP procedures — achieving more than 80% gain in processing speed while making the logic markedly more flexible to extend.