jorgecuadros-platform/scripts/numid-audit.sql

-- One row per portal NUMid, with every signal that says whether the id is in use.
-- Consumed by scripts/numid-audit.mjs, which applies the tier rules.
--
-- POOL. customer_legacy_refs where sourceSystem='utilities' AND sourceTable='DATGRAL'.
-- That pair IS the portal "Security Number" the login screen asks for.
-- insurance/DATGRAL is a DIFFERENT id space running to 4000 and sharing the same
-- sourceTable name; drawing from it would hand out an id the portal cannot resolve.
--
-- WHY THE OBVIOUS RULES FIND NOTHING.
--   "every owned row count is zero"      -> 0 of 1,171. Migration gave every NUMid
--                                           at least one property and one transaction.
--   "no transaction in the last N years" -> 0 of 1,171. Every customer carries a
--                                           synthetic Jan-1 opening-balance row, so
--                                           everyone looks active in the current year.
-- The opening-balance row has to be subtracted before any of this means anything,
-- which is what `bf` below does and why `real_tx` exists.
--
-- BALANCE-FORWARD DETECTION IS TWO-SHAPED ON PURPOSE.
-- transform_transactions.py:120 mints a transaction type literally named
-- 'BALANCE FORWARD'. Databases loaded before that change carry the same rows with
-- typeId NULL, dated Jan 1, legacySourceTable='datos2' -- 1,170 of them, exactly one
-- per customer. Matching the type name alone floors nothing on such a database, and
-- every balance below silently becomes a raw lifetime sum: the same double-count that
-- read the whole book as +20.6M MXN in credit before d173c9e. Match both shapes.
--
-- Balances otherwise follow BillingService exactly -- voided out, outstanding out,
-- superseded rows out (BALANCE_FLOOR_JOIN / NOT_SUPERSEDED, billing.service.ts:179-210).

WITH bf AS (
  SELECT t.id, t.customerId, t.transactionDate
  FROM transactions t
  LEFT JOIN type_transactions tt ON tt.id = t.typeId
  WHERE t.voidedAt IS NULL
    AND (
      tt.nameEn = 'BALANCE FORWARD'
      OR (t.typeId IS NULL AND MONTH(t.transactionDate) = 1 AND DAY(t.transactionDate) = 1
          AND t.legacySourceTable = 'datos2')
    )
),
bfloor AS (
  SELECT customerId, MAX(transactionDate) AS floorDate FROM bf GROUP BY customerId
),
real_tx AS (
  SELECT t.* FROM transactions t
  WHERE t.voidedAt IS NULL AND t.id NOT IN (SELECT id FROM bf)
),
pool AS (
  SELECT CAST(r.legacyId AS UNSIGNED) AS numid,
         c.id AS cid,
         REPLACE(REPLACE(COALESCE(c.name,''),'\n',' '),'\t',' ') AS name,
         IF(c.archivedAt IS NULL,0,1) AS archived,
         IF(c.email IS NULL OR c.email='',0,1) AS hasEmail
  FROM customer_legacy_refs r
  JOIN customers c ON c.id = r.customerId
  WHERE r.sourceSystem='utilities' AND r.sourceTable='DATGRAL'
)
SELECT
  p.numid, p.cid AS customerUuid, p.name, p.archived, p.hasEmail,
  -- EXISTS, not a join: 16 customers hold more than one insurance ref (several
  -- insurance rows folded into one customer), and joining them fans this result
  -- out past one row per NUMid — 1,188 rows for a 1,171-id pool.
  EXISTS(SELECT 1 FROM customer_legacy_refs i
         WHERE i.customerId=p.cid AND i.sourceSystem='insurance') AS insRef,

  COALESCE((SELECT ROUND(SUM(t.amount),2) FROM transactions t
            LEFT JOIN bfloor f ON f.customerId=t.customerId
            WHERE t.customerId=p.cid AND t.voidedAt IS NULL AND t.outstanding=0
              AND t.currency='MXN'
              AND (f.floorDate IS NULL OR t.transactionDate>=f.floorDate)),0) AS balMxn,
  COALESCE((SELECT ROUND(SUM(t.amount),2) FROM transactions t
            LEFT JOIN bfloor f ON f.customerId=t.customerId
            WHERE t.customerId=p.cid AND t.voidedAt IS NULL AND t.outstanding=0
              AND t.currency='USD'
              AND (f.floorDate IS NULL OR t.transactionDate>=f.floorDate)),0) AS balUsd,

  (SELECT COUNT(*) FROM transactions t
     WHERE t.customerId=p.cid AND t.voidedAt IS NULL AND t.outstanding=1) AS nopago,
  (SELECT COUNT(*) FROM real_tx t WHERE t.customerId=p.cid) AS realTx,
  (SELECT COUNT(*) FROM real_tx t WHERE t.customerId=p.cid
     AND t.transactionDate >= DATE_SUB(CURDATE(), INTERVAL 12 MONTH)) AS realTx12m,
  (SELECT COUNT(*) FROM real_tx t WHERE t.customerId=p.cid
     AND t.transactionDate >= DATE_SUB(CURDATE(), INTERVAL 36 MONTH)) AS realTx36m,
  (SELECT DATE(MAX(t.transactionDate)) FROM real_tx t WHERE t.customerId=p.cid) AS lastRealTx,

  (SELECT COUNT(*) FROM properties pr WHERE pr.customerId=p.cid AND pr.archivedAt IS NULL) AS props,
  -- services are counted BOTH ways: an inactive service is still a record of the id
  -- having been used, so the auto tier requires zero of any kind.
  (SELECT COUNT(*) FROM property_services ps JOIN properties pr ON pr.id=ps.propertyId
     WHERE pr.customerId=p.cid AND pr.archivedAt IS NULL) AS anySvc,
  (SELECT COUNT(*) FROM property_services ps JOIN properties pr ON pr.id=ps.propertyId
     WHERE pr.customerId=p.cid AND pr.archivedAt IS NULL AND ps.active=1) AS activeSvc,
  (SELECT COUNT(*) FROM policies po WHERE po.customerId=p.cid AND po.archivedAt IS NULL
     AND (po.policyTo IS NULL OR po.policyTo >= CURDATE())) AS activePol,
  (SELECT COUNT(*) FROM policies po WHERE po.customerId=p.cid AND po.archivedAt IS NULL) AS anyPol,
  (SELECT COUNT(*) FROM vehicles v WHERE v.customerId=p.cid) AS veh,
  (SELECT COUNT(*) FROM trust_accounts ta JOIN properties pr ON pr.id=ta.propertyId
     WHERE pr.customerId=p.cid) AS trust,
  (SELECT COUNT(*) FROM statement_documents s WHERE s.matchedCustomerId=p.cid) AS stmt,
  (SELECT COUNT(*) FROM policy_ocr_documents o WHERE o.matchedCustomerId=p.cid) AS ocr,
  (SELECT COUNT(*) FROM email_notification_log e WHERE e.customerId=p.cid) AS enl,
  (SELECT COUNT(*) FROM email_log e WHERE e.customerId=p.cid) AS elog,
  (SELECT COUNT(*) FROM account_status_history a WHERE a.customerId=p.cid) AS ash
FROM pool p
ORDER BY p.numid;