Skip to content

feat(db): Onda 19 — precisão explícita em 21 colunas numeric (item 5.4) #214

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
adm01-debug merged 1 commit into from
May 15, 2026
Merged

feat(db): Onda 19 — precisão explícita em 21 colunas numeric (item 5.4) #214

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
156 changes: 156 additions & 0 deletions supabase/migrations/20260515030000_onda19_numeric_precision.sql
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,156 @@
-- =================================================================
-- Onda 19: Precisão explícita em colunas numeric (audit gap 5.4)
--
-- Contexto:
-- Auditoria de 10/mai/2026 (item 5.4) mapeou 21 colunas `numeric`
-- sem precisão definida — escala ilimitada permite drift silencioso
-- (ex: 1234.567899 em coluna de dinheiro, frontend exibe 2 casas
-- mas DB guarda original; comparações de igualdade ficam frágeis).
--
-- Banco em estado pré-prod (~0 linhas em todas afetadas, exceto
-- quotes.real_subtotal com 3 linhas escala=2). Zero drift detectado
-- na varredura pré-aplicação.
--
-- Padronização adotada (alinhada com padrão existente do projeto):
-- - Dinheiro principal: numeric(10,2) (até R$ 99.999.999,99)
-- - Dinheiro grandes totais: numeric(12,2) (até R$ 9.999.999.999,99 — kits, favoritos)
-- - Percentuais 0-100: numeric(5,2)
-- - Markup (pode > 100%): numeric(5,2) (até 999,99% — decisão A do PO)
-- - Confidence (0-1): numeric(3,2)
-- - Rate fiscal (% inteiro): numeric(5,2) (alinha com icms_rate; ipi até 20%)
-- - Dimensões cm/cm²: numeric(8,2)
--
-- Dependências resolvidas (DROP + ALTER + RECREATE no mesmo migration):
-- - VIEW v_audit_paradoxos_gravacao (referencia markup_percent)
-- - TRIGGER trg_quotes_calc_real_values (UPDATE OF negotiation_markup_percent)
-- - TRIGGER trg_kit_print_area_normalizar_eixos (UPDATE OF max_width, max_height)
--
-- NÃO MEXIDO (decisão pré-flight):
-- - orders.total (GENERATED ALWAYS AS total_amount — herda precisão)
-- - app_vitals.metric_value (telemetria livre)
-- - _asia_api_staging.peso, color_analysis_staging.match_distance (staging)
-- - tabela_preco_gravacao_oficial.{preco_max,preco_min}_unitario (placeholder)
-- =================================================================

-- 0. DROP dependências bloqueadoras
DROP VIEW IF EXISTS public.v_audit_paradoxos_gravacao;
DROP TRIGGER IF EXISTS trg_quotes_calc_real_values ON public.quotes;
DROP TRIGGER IF EXISTS trg_kit_print_area_normalizar_eixos ON public.kit_component_print_areas;
Comment on lines +35 to +38

-- 1. PERCENTUAIS (8 colunas)
ALTER TABLE public.quotes
ALTER COLUMN discount_percent TYPE numeric(5,2),
ALTER COLUMN negotiation_markup_percent TYPE numeric(5,2),
ALTER COLUMN real_discount_percent TYPE numeric(5,2);

Comment on lines +42 to +45
ALTER TABLE public.seller_discount_limits
ALTER COLUMN max_discount_percent TYPE numeric(5,2),
ALTER COLUMN approval_required_above TYPE numeric(5,2);

ALTER TABLE public.tabela_preco_gravacao_oficial
ALTER COLUMN markup_percent TYPE numeric(5,2);

ALTER TABLE public.supplier_technique_mappings
ALTER COLUMN confidence TYPE numeric(3,2);

ALTER TABLE public.variant_supplier_sources
ALTER COLUMN supplier_ipi_rate TYPE numeric(5,2);

-- 2. DINHEIRO (8 colunas)
ALTER TABLE public.quotes
ALTER COLUMN real_subtotal TYPE numeric(10,2);

ALTER TABLE public.quote_item_personalizations
ALTER COLUMN unit_cost TYPE numeric(10,2),
ALTER COLUMN setup_cost TYPE numeric(10,2),
ALTER COLUMN total_cost TYPE numeric(10,2);

ALTER TABLE public.kit_variants
ALTER COLUMN total_price TYPE numeric(12,2);

ALTER TABLE public.seller_cart_items
ALTER COLUMN product_price TYPE numeric(10,2);

-- collection_items + trash mantêm paridade (favorite_items.price_at_save já é numeric(12,2))
ALTER TABLE public.collection_items
ALTER COLUMN price_at_save TYPE numeric(12,2);

ALTER TABLE public.collection_items_trash
ALTER COLUMN price_at_save TYPE numeric(12,2);

-- 3. DIMENSÕES (5 colunas)
ALTER TABLE public.quote_item_personalizations
ALTER COLUMN area_cm2 TYPE numeric(8,2),
ALTER COLUMN height_cm TYPE numeric(8,2),
ALTER COLUMN width_cm TYPE numeric(8,2);

ALTER TABLE public.kit_component_print_areas
ALTER COLUMN max_height TYPE numeric(8,2),
ALTER COLUMN max_width TYPE numeric(8,2);

-- 4. RECREATE triggers (definições idênticas às originais)
CREATE TRIGGER trg_quotes_calc_real_values
BEFORE INSERT OR UPDATE OF subtotal, discount_amount, negotiation_markup_percent
ON public.quotes
FOR EACH ROW
EXECUTE FUNCTION public.fn_quotes_calc_real_values();
Copy link
Copy Markdown

@chatgpt-codex-connector chatgpt-codex-connector Bot May 15, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Badge Restore the checked-in quote trigger function

When this migration is applied from a clean database built from the repository migrations, public.fn_quotes_calc_real_values() is never created (repo-wide search of supabase/migrations and recovery only finds this call), so PostgreSQL will fail at CREATE TRIGGER ... EXECUTE FUNCTION and roll back the migration before any later steps run. Either recreate the trigger against the existing checked-in function (validate_quote_real_discount) or add/commit the missing function before this trigger is created.

Useful? React with 👍 / 👎.


CREATE TRIGGER trg_kit_print_area_normalizar_eixos
BEFORE INSERT OR UPDATE OF max_width, max_height
ON public.kit_component_print_areas
FOR EACH ROW
EXECUTE FUNCTION public.fn_kit_print_area_normalizar_eixos();
Copy link
Copy Markdown

@chatgpt-codex-connector chatgpt-codex-connector Bot May 15, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Badge Restore the checked-in kit trigger function

The same clean-migration path fails here because public.fn_kit_print_area_normalizar_eixos() is not defined anywhere in the checked-in migrations or recovery SQL; CREATE TRIGGER requires the referenced trigger function to already exist. This makes supabase db reset/new environment provisioning depend on untracked production drift instead of the repository state.

Useful? React with 👍 / 👎.


-- 5. RECREATE view (definição idêntica à original — só rebuild para usar novo tipo)
CREATE OR REPLACE VIEW public.v_audit_paradoxos_gravacao AS
Copy link
Copy Markdown

@chatgpt-codex-connector chatgpt-codex-connector Bot May 15, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Badge Preserve the audit view hardening

docs/redeploy/REDEPLOY-FASE2-EXECUTION-LOG.md classifies v_audit_paradoxos_gravacao as internal/service-role-only and says those views were hardened with security_invoker=true plus role revokes; dropping and recreating it here does not restore either setting. In environments with the same exposed-schema/default grants, this silently reverts the view to owner-rights semantics and can expose the underlying pricing audit rows without the intended RLS/role boundary, so the migration should reapply the ALTER VIEW ... SET (security_invoker = true) and grants/revokes after creation.

Useful? React with 👍 / 👎.

WITH faixas_ord AS (
SELECT t.id,
t.codigo_tabela,
t.nome,
t.grupo_tecnica,
t.custo_setup,
t.markup_percent,
f.quantidade_minima,
f.quantidade_maxima,
f.preco_unitario,
lag(f.preco_unitario) OVER (
PARTITION BY t.id,
(COALESCE(f.largura_min, 0::numeric)),
(COALESCE(f.largura_max, 0::numeric)),
(COALESCE(f.altura_min, 0::numeric)),
(COALESCE(f.altura_max, 0::numeric))
ORDER BY f.quantidade_minima
) AS preco_anterior,
lag(f.quantidade_maxima) OVER (
PARTITION BY t.id,
(COALESCE(f.largura_min, 0::numeric)),
(COALESCE(f.largura_max, 0::numeric)),
(COALESCE(f.altura_min, 0::numeric)),
(COALESCE(f.altura_max, 0::numeric))
ORDER BY f.quantidade_minima
) AS qty_max_anterior
FROM public.tabela_preco_gravacao_oficial t
JOIN public.tabela_preco_gravacao_oficial_faixa f
ON f.tabela_preco_gravacao_id = t.id
WHERE t.ativo = true
)
SELECT
codigo_tabela,
nome,
grupo_tecnica,
qty_max_anterior AS qty_pico_ant,
preco_anterior,
round(GREATEST(qty_max_anterior::numeric * preco_anterior, custo_setup) * (1::numeric + markup_percent / 100::numeric), 2) AS venda_no_pico,
quantidade_minima AS qty_inicio,
preco_unitario AS preco_atual,
round(GREATEST(quantidade_minima::numeric * preco_unitario, custo_setup) * (1::numeric + markup_percent / 100::numeric), 2) AS venda_inicio,
round(GREATEST(quantidade_minima::numeric * preco_unitario, custo_setup) - GREATEST(qty_max_anterior::numeric * preco_anterior, custo_setup), 2) AS economia_cliente_se_subir_faixa,
CASE
WHEN preco_anterior IS NULL THEN 'primeira_faixa'::text
WHEN qty_max_anterior IS NULL THEN 'sem_pico_anterior'::text
WHEN GREATEST(quantidade_minima::numeric * preco_unitario, custo_setup) < GREATEST(qty_max_anterior::numeric * preco_anterior, custo_setup) THEN 'PARADOXO_NATURAL'::text
ELSE 'OK'::text
END AS status_natural
FROM faixas_ord
WHERE preco_anterior IS NOT NULL AND qty_max_anterior IS NOT NULL
ORDER BY codigo_tabela, quantidade_minima;
Loading