SAP S/4HANA performance optimization: project method

In SAP projects, performance is not a comfort topic. It drives process reliability, SLA compliance, and the overall perception of the solution by the business. In technical interviews, it often separates a decent developer from a senior consultant.

In S/4HANA, implementation choices show immediately: HANA amplifies good decisions (pushdown, targeted queries) and punishes bad ones (loops, SELECT * and unnecessary ABAP processing).

Impacts are measurable: batch duration, response time, scalability, infrastructure cost, and error rates. Technically, this translates to poorly targeted DB queries, oversized internal tables, and overly verbose code paths.

Golden rule in S/4HANA: let the database work. Pushdown through CDS Views and modern Open SQL reduces data volume, network latency, and application load.

Recommended flow
User
  |  (UI filters)
  v
CDS View (projection)
  |  (associations, annotations)
  v
CDS View (interface)
  |  (joins, calculations)
  v
HANA database
  |  (pushdown, aggregations)

"ABAP 7.50+ : targeted Open SQL and pushdown
SELECT FROM zso_hdr AS hdr
  INNER JOIN zso_itm AS itm
    ON itm~so_id = hdr~so_id
  FIELDS hdr~so_id,
         hdr~customer,
         SUM( itm~net_amount ) AS net_total
  WHERE hdr~status = @lv_status
    AND hdr~created_on BETWEEN @lv_date_from AND @lv_date_to
  GROUP BY hdr~so_id, hdr~customer
  INTO TABLE @DATA(lt_result).

In a real project, this simple pushdown divided a sales monitoring screen by 6 by replacing ABAP processing with SQL aggregation.

• Limit columns and rows from the SELECT.
• Prefer JOINs, GROUP BY, and HAVING on the DB side.
• Check indexes and selectivity of conditions.

ABAP 7.50+ provides cleaner syntax, but performance depends on internal table structure, sorting, and reducing loops.

"Internal table suited to the use case
TYPES: BEGIN OF ty_line,
         so_id     TYPE zso_id,
         customer  TYPE kunnr,
         net_total TYPE p LENGTH 16 DECIMALS 2,
       END OF ty_line.

DATA lt_data TYPE HASHED TABLE OF ty_line WITH UNIQUE KEY so_id.

Using a HASHED TABLE with a unique key avoids O(n) READ TABLE and secures lookups.

• Avoid SELECT inside loops; group reads instead.
• Use READ TABLE ... BINARY SEARCH on sorted tables.
• Reduce volume before any ABAP processing.

In batch processing, memory often becomes the bottleneck. An oversized internal table can exhaust work process memory and trigger dumps.

• Explicitly FREE large tables after use.
• Avoid unnecessary copies (MOVE-CORRESPONDING on big volumes).
• Load in packages (PACKAGE SIZE) for large extractions.

"Extraction in packages
SELECT FROM zso_hdr
  FIELDS so_id, customer, net_total
  INTO TABLE @DATA(lt_hdr)
  PACKAGE SIZE 5000.

  "Processing
  CLEAR lt_hdr.
ENDSELECT.

Analysis is tool-driven, not intuition-driven. Start by isolating the symptom, then refine with the right tools.

Quick analysis workflow
User signal
  |-> SAT (ABAP profile)
  |-> ST05 (SQL trace)
  |-> SM50/SM66 (WP)
  |-> ST22 (dump)
  v
Summary + action plan

Here are realistic scenarios encountered on projects, with a reproducible resolution approach.

• SELECT * when 5 fields are enough.
• SELECT inside a loop without caching.
• FOR ALL ENTRIES without checking empty tables.
• Sorting in ABAP when ORDER BY would be enough.
• Standard internal tables for heavy lookups.

These anti-patterns remain the main causes of red ST05 and saturated SAT traces.

Performance is a pillar of Clean Core: less custom code, more standard, more published APIs. In S/4HANA this means ATC, Released APIs, and ABAP Cloud compliance.

• Limit standard modifications to remain upgrade compatible.
• Use CDS Views and SAP-exposed services.
• Validate continuously with ATC + SCI.

It is also a strong interview argument: you show that performance is an architecture topic, not just local tuning.

• Start by measuring: SAT and ST05 before any optimization.
• Push calculation to HANA with CDS and modern Open SQL.
• Choose the right internal table structure for the use case.
• Document gains and secure with ATC/SCI.