CORREGIR 10: importador debe ingerir altas recientes y filas cerradas

Sub-tickets

Activity

• wi-cli-venus created · 2026-08-14
  parent=#2590
• wi-cli-venus commented · 2026-08-20
  MOVILBA-63 canceled into this: tools/rutap-sync-prod.sql must exclude prod rows already in estado BAJA before concluding a dominio has no match in the book. Without that predicate every run re-manufactures the same false positive - 130 closed-period rows read as stale because the book's open row describes the NEW plate, which prod also holds. One predicate, not a new detector.
• wi-cli-venus completed · 2026-08-20
  EXECUTED on PROD 2026-08-20 with Elazar's GO, COMMIT, verified after. Written exactly as rehearsed: pass A 2 vdc rows, pass B 3 vdc + 3 vehiculo_historial. Quarantine 1 / resolved 1 (AH924PV 1900-01-06 -> 2026-04-15, derived from 11 siblings on EX-2026-17669949, 1 distinct date). Gates all zero: 0 vigentes absent from prod, 0 with more than one open row, 0 closed without expte_baja, 0 with an ambiguous last baja, 0 suspicious chasis. AFTER: AB674IP 2026-08-10 / EX-2026-36096231 and AF106VS 2026-08-07 / EX-2026-35826642, both matching the book; AH924PV unchanged at 2026-04-15 because the derived value already agreed with prod - the derivation self-checked by writing nothing. Export totals unchanged 1736 = 1599 Alta + 137 Baja: the sync moved no vehicle between states, it corrected tramite dates and expedientes on rows that stay Alta. tools/rutap-sync-checks.sql green post-COMMIT: 137 closed periods retained with borrado_en NULL, 130 replate groups still 1 ALTA + n BAJA, 0 chasis twice in Alta. Ships with: date quarantine + derivation at n>=3, chasis identity pre-pass, chasis cross-check in pass 0, 5-class contamination scan, replate fixture.