Module 18: Code Pushdown Concept

Code pushdown means moving data-intensive logic from the ABAP application server to the database layer.
With SAP HANA’s in-memory architecture, pushdown is a core performance principle, not an optimization afterthought.

---

1. Why Code Pushdown?

Traditional ABAP Model (Pre-HANA)

• Database → Data retrieval only
• Application server → Heavy processing
• Large data transfers

HANA Model

• Database → Data retrieval + processing
• Application server → Orchestration & business logic

HANA Philosophy

Move data-intensive logic to where the data lives.

---

2. DB vs Application Server Logic

---

Application Server Logic (Classic)

SELECT * FROM zsales
  INTO TABLE @DATA(lt_sales).

LOOP AT lt_sales INTO DATA(ls_sale).
  IF ls_sale-amount > 1000.
    ls_sale-category = 'HIGH'.
  ELSE.
    ls_sale-category = 'LOW'.
  ENDIF.
ENDLOOP.

Database Pushdown (Modern)

SELECT *,
       CASE
         WHEN amount > 1000 THEN 'HIGH'
         ELSE 'LOW'
       END AS category
  FROM zsales
  INTO TABLE @DATA(lt_sales).

Result

Less data transfer, fewer loops, better performance. ::

3. Pushdown Options in ABAP

Technique	Pushdown Level
Open SQL (advanced)	Medium
CDS Views	High
CDS Table Functions	Very High
AMDP	Maximum

SAP Recommendation

Prefer Open SQL → CDS → AMDP in that order.

4. Performance Comparison

Scenario: 1 Million Records

Approach	Data Transfer	App Server Load	Performance
Loop-based ABAP	High	High	Poor
Open SQL expressions	Medium	Low	Good
CDS View	Low	Very Low	Excellent
AMDP	Minimal	Minimal	Best

Interview Trap

Pushdown is not about writing SQL everywhere — it is about placing logic correctly.

5. What Should Be Pushed Down?

Good Candidates

• Filtering

• Aggregations

• Calculations

• Joins

• Grouping

-Sorting

SELECT carrid,
       SUM( price ) AS total_price
  FROM zflight
  GROUP BY carrid
  INTO TABLE @DATA(lt_totals).

6. What Should NOT Be Pushed Down?

6.1 Complex Business Logic

❌ Poor Pushdown Candidate

IF user_role = 'ADMIN' AND system_state = 'CLOSED'.

Reason:

• Depends on application context

• Hard to express in SQL

• Poor readability

6.2 UI & Authorization Logic

• UI behavior

• User messages

• Authorization checks

Separation of Concerns

Database = data logic Application server = business & orchestration logic

6.3 Small Data Volumes

Rule of Thumb

If dataset is small, clarity > pushdown.

7. When NOT to Push Down (Clear Rules)

Do NOT Push Down When

• Logic is procedural or stateful

• Code becomes unreadable

• Business rules change frequently

• You need reuse across DBs

• Performance gain is negligible

8. Code Pushdown Anti-Patterns

Avoid These

• Moving everything into SQL

• Writing unreadable monster queries

• Using Native SQL unnecessarily

• Ignoring maintainability

• Overusing AMDP for simple logic

9. Interview-Grade Answer Template

Q: When should code be pushed down to HANA?

Answer:

Push down logic when it is data-intensive, set-based, and independent of application context. Avoid pushdown for complex business logic, UI behavior, or small datasets. Prefer Open SQL → CDS → AMDP based on complexity.

Use This Answer

This framing is exactly what interviewers want to hear.

10. Summary

• Pushdown improves performance by reducing data transfer

• HANA enables in-memory processing

• Not all logic should be pushed down

• Choose the right pushdown technique

• Maintainability matters as much as performance

11. Practice Exercises

• Convert loop-based logic into SQL CASE expression.

• Replace ABAP aggregation with GROUP BY.

• Identify pushdown candidates in a report.

• Decide when NOT to push down.

• Compare Open SQL vs CDS pushdown.

12. What’s Next?

➡️ Module 19: CDS Views – Fundamentals

Learning Tip

• Good developers push code down.
• Great developers know when not to.