database-hana

SAP HANA Database Standards

Priority: P0 (CRITICAL)

Rules

• Strict Parameterization: Always use ? parameter placeholders for SAP HANA queries. Never use raw string concatenation, fmt.Sprintf, or untrusted interpolation for values.
• Batching & Chunking Limits: Chunk large IN (...) parameter lists into batches of at most 1,000 items to avoid HANA driver and engine parameter limits.
• Explicit Column Aliasing: Always explicitly alias columns in multi-table queries (e.g. m.code AS code, m.name AS name). SAP HANA and Go database drivers map scan columns by name; duplicate column names across joins silently collide or overwrite values. Never use SELECT *.
• HANA Datatype Casting Gotchas: SAP HANA functions (such as COALESCE, SUBSTR, CONCAT) require strictly compatible parameter types. Explicitly cast values when comparing or falling back (e.g. COALESCE(col, 'default'), TO_VARCHAR(...), CAST(? AS NVARCHAR)).
• Identifier Case-Sensitivity: Unquoted SQL identifiers in SAP HANA default to UPPERCASE. Quoted identifiers ("columnName") preserve case. Keep identifiers consistent with schema conventions.
• Keyset & Streaming Pagination: Avoid deep LIMIT/OFFSET on large columnar tables. Prefer keyset/seek pagination (WHERE id > ? ORDER BY id ASC LIMIT ?) and stream large result sets using rows.Next() directly into models to prevent memory spikes.
• Transaction Scope: Keep transactions short and focused on atomic write operations. Never perform slow external I/O or unbounded operations inside an open transaction. Always pair with defer tx.Rollback().

Verify

• Query uses ? placeholders for all user-supplied or dynamic values.
• Large IN (...) queries chunk parameters into batches of ≤ 1,000 items.
• All joined queries use explicit, distinct column aliases without SELECT *.
• Built-in functions (COALESCE, SUBSTR) use matching datatypes or explicit CAST/TO_VARCHAR.
• Pagination enforces strict LIMIT caps and stable ORDER BY.
• Transactions are scoped to write paths with defer tx.Rollback().