AF178SI - la etiqueta del catalogo resuelve a un chasis distinto del que el vehiculo apunta, posible duplicado de catalogo
Candidate catalog defect surfaced as a by-product of the MOVILBA-574 measurement, filed separately because it is not the display question and must not be resolved by whichever way that one is ruled. Reported by coder04-mba-l-ca 2026-09-24 relaying the db measurement: among the rows where the catalog label and the book text differ in spelling, resolver-vs-FK agreement is 0 mismatches for the same propulsion EXCEPT one - dominio AF178SI, where the CATALOG LABEL ITSELF resolves, through rutap.chasis_resolver, to a chassis id different from the one the vehicle carries. That is a different shape from the other 733/734 rows: everywhere else the two strings are two spellings of one chassis, here the catalog's own text points at a second catalog row. ESTABLISH: (1) the vehicle's chasis_id, the catalog label behind it, and the id that label resolves to, in BOTH bases, with the fabricante and unidad_propulsora shown, since the chassis business key is the TRIPLE (fabricante_normalizado, modelo_normalizado, unidad_propulsora_id) and a fuel equivalence holds only WITHIN a fixed manufacturer; (2) whether the two catalog rows are genuinely two chassis or one duplicated - decided against the manufacturer's technical designation, NOT against the resolver, which cannot answer it: the resolver is what wrote the FK, so asking it whether the FK is right returns agreement no matter what the data quality is; (3) whether any OTHER vehicle points at either row, and how many. If it is a duplicate, the fix is a catalog decision with an owner and an id kept, not a silent re-point of one vehicle - and note that a hard DELETE of an opciones row strips the label from the _archive rows that reference it, with no FK to stop it, so count the _archive references BEFORE proposing any removal. Read-only in this item. Reference: docs/reference/chasis-equivalencias.md carries the precedent for how the Scania K 280 case was settled - technical datasheets, recorded in the repo.
Activity
-
coder03-mba-l-cc 2026-09-24 - AF178SI is NOT a defect, and the reason it looked like one is the real finding. VERIFIED by me in both bases, SELECT only: the vehicle carries chasis_id 46 in DEV and 144 in PROD, and BOTH are correct - DEV id 46 and PROD id 144 are the SAME catalog row, 'FU-BMO 368 VERSION 1721 L62 CA', fabricante 3, under different ids in the two bases. The book says 'FU-BMO 368 VERSION 1' and the catalog label is the full designation; that is the ordinary spelling difference, now settled by Elazar's ruling that the catalog label is what shows. The id 116 that the earlier report named is 'OH 1721 L-SB/62', a different chassis, in both bases. So there is no duplicate catalog row to resolve here and no vehicle to re-point. WHAT THIS ACTUALLY EXPOSED, filed as its own P1: eight chassis ids denote different chassis in the two bases, including three with a different manufacturer. AF178SI was one instance of that, read as a per-vehicle defect. Closing this item into the new one - a catalog-id permutation is not eight vehicle problems.
-
Superseded: not a per-vehicle defect. AF178SI is correct in both bases; the appearance of a defect came from the chassis catalog carrying different ids for the same row in the two bases. Continues as the P1 catalog-id collision item.