Rule of Thumb: DB First

Explanation

Most performance problems are database problems. Aggregate in SQL (SUM/GROUP BY), avoid SELECT * (project explicit columns), use INNER JOINs over nested SELECTs, and consider CDS or AMDP for HANA-native processing. Only tune ABAP loops after the SQL is confirmed efficient.

Code example

SELECT bukrs, SUM( dmbtr ) AS total  FROM bkpf  WHERE budat BETWEEN @lv_from AND @lv_to  GROUP BY bukrs  INTO TABLE @DATA(lt_totals).

Real project scenario

Aggregating in SQL cut a 900k-row report from 15 minutes to 4 seconds — no ABAP change needed.

Common mistakes

SELECT * INTO INTERNAL TABLE then LOOP + SUM on the app server; FOR ALL ENTRIES with duplicates.

Best practices

Explicit columns; join in DB when possible; measure before and after with ST05.

Interview angle

When does aggregating in ABAP still make sense?