Lesson 3 of 6
Assessing flows, routines and queries
9 min
Listen to this lesson
Before Contoso chooses a path per use case, Noor needs an inventory of BWP: what exists, what is used, and what is risky to move. Contoso's conversion program did the same for custom code (11,800 Z objects, 4,300 deleted as unused). BW deserves the same discipline.
What to inventory
| Object | Where to find it | What to record |
|---|---|---|
| Queries | Query definitions; BW statistics (query runtime and usage data, for example from the BW technical content or the RSDDSTAT tables) | Usage in the last 12 months, users, criticality (close, regulatory), complexity |
| InfoProviders | Data Warehousing Workbench (RSA1) or BW modeling tools | Type (InfoCube, DSO, aDSO, CompositeProvider), size, consumers |
| Data flows and DataSources | Data flow display; source system DataSources | Source (CMP, flat file, non-SAP), extractor support in S/4HANA |
| Transformations and routines | Transformation definitions; ABAP in start, end, expert and field routines | Lines of code, lookups, hard-coded values, HANA-optimization |
| Process chains | Process chain maintenance (RSPC) | Schedule, dependencies, failure history |
| Authorizations | Analysis authorizations (RSECADMIN) | Which restrictions protect which data |
A BW inventory. The usage column decides scope; the routine and authorization columns decide risk.
Scoring migration risk
Each use case (a set of queries with the providers and flows behind them) gets a usage score and a risk score. Usage comes from BW statistics: number of executions and users in the last 12 months. Risk comes from the objects behind it: ABAP routines (especially lookups and expert routines), extractors not supported in S/4HANA, classic objects that need conversion, complex query features (exception aggregation, conditions, hierarchies, currency conversion), and analysis authorizations. Noor's illustrative results for BWP: about 1,400 queries, 40% with no use in 12 months, 85 use cases in scope, 14 of them with expert routines.
| Use case | Usage (12 months) | Risk drivers | Proposed path |
|---|---|---|---|
| Monthly margin by customer (finance) | High: 40 users, every close | End routine for cost lookup by plant; currency translation | Lift, expose, then rebuild in Datasphere with regression |
| Sales by channel (O2C) | High: Maria's team weekly | Billing extractor 2LIS_13_VDITM | Replace with SAP-managed data product if content matches |
| Core return agreements report | Medium | Custom DataSource on the Z table carried over from ECC | Rebuild on a CDS-based extraction |
| Old plant 1720 capacity report | None in 12 months | Expert routine | Retire |
| Inventory turns | Medium | Snapshot InfoCube, classic object | Lift; rebuild later with snapshots in Datasphere |
An extract of Contoso's BW use case assessment. Figures are illustrative course data.
Authorizations need their own line in the assessment. BW analysis authorizations in RSECADMIN restrict data by characteristic values, for example a controller who sees only company code 1810 or only profit center CMC-PARTS. None of that travels automatically to Datasphere, SAC or SAP Databricks. For every use case that moves, Nadia’s team maps the BW restrictions to Datasphere data access controls or SAC data access, and Priya adds negative access tests to the regression set. A rebuild that reconciles perfectly but shows 1810 data to an 1710-only controller has failed.
What would you do?
The inventory shows 560 queries with no use in 12 months. The sales team says some "might be needed for year-end". What should Noor do?
- Extract usageBW statistics for 12+ months, including a year-end
- Group into use casesQueries with their providers, flows and chains
- Score riskRoutines, extractors, classic objects, query features, authorizations
- Confirm with ownersDan, Maria and other owners confirm retire or keep
- Assign a pathRetire, lift, expose, rebuild or convert, with a test approach
Process chains are the last part of the inventory, and the one most often forgotten. A chain encodes order and timing: master data before transaction data, deltas before snapshots, the inventory snapshot after the last goods movement posting of the month. When use cases move, those dependencies must be rebuilt as task chains or schedules in Datasphere, or the new model will be right on test data and wrong on the first month-end. Noor lists every chain with its schedule, dependencies and failure history, and Priya adds a month-end rehearsal to the regression plan for any use case with period-end snapshots.
- Inventory queries, providers, flows, routines, process chains and analysis authorizations.
- Score usage from BW statistics over a full year including year-end; score risk from routines, extractors and query features.
- Routines hide business rules: document each and turn it into a regression case.
- Retire unused content with owner confirmation, and keep definitions for reference.
Check your understanding · not graded
Lesson check
1. Where does query usage evidence come from?
2. Why are ABAP routines a high migration risk?
3. Why should usage be measured over a window that includes year-end?
4. Which transaction maintains BW analysis authorizations?