Skills

development/performance/ifs-mass-data-insert-scripts

ifs-mass-data-insert-scripts

Delivers simple Oracle PL/SQL that bulk-duplicates seed rows into IFSAPP tables using a plain BEGIN/FOR LOOP, INSERT … SELECT … WHERE seed pattern, batched COMMIT, SYS_GUID() for ROWKEY where needed, and parent-before-child FK order with matching key transforms (“relevant connection”) between clones. Generated runnable scripts contain only INSERT-driven duplication (seeded SELECT) plus COMMIT and block syntax—and **standalone preflight SELECT/INTO/COUNT** where seed chains require it— never DELETE/UPDATE/MERGE, never DDL (create/drop/alter/truncate tables or schema). Test-plan mode follows **three gates (MUST RULE)**: (1) **deep discovery** — walk the entire plan + codebase and identify **every** persisted table used in the logic before listing anything; (2) **table inventory + performance impact % + requirements Q&A** — present the full list with estimated per-table performance impact, then ask the user to finalize volumes, seeds, scope, and exclusions; (3) **script creation** — only after Gates 1–2 are satisfied. Includes manufacturing coherence (BOM, routing, costing) for PRODUCT-like flows. Use for mass inserts, synthetic volume data, bulk duplication scripts, test-plan-driven volume, or invokes ifs-mass-data-insert-scripts. Scripts must compile when seeds exist; prefer static SQL unless the user explicitly asks otherwise. Before emitting INSERT lists, deep-scan related *.entity files (and BASED_ON, storage, upg) for mandatory columns, keys, and constraints, then reconcile with all_tab_columns on the target DB. Every INSERT must include ROWVERSION and ROWKEY when catalog marks them NOT NULL (listed last; seed.rowversion + SYS_GUID()).