Referential integrity with explicit exceptions

The rules

These rules decide what a correct result is. Your answer is checked on this dataset and on hidden datasets that follow the same rules.

Inputs
  payments(id, account, amount_cents) — text. id is unique. amount_cents is always a valid whole number (refunds are negative).
  accounts(account, name) — text. Accounts are trimmed and unique, never missing.

Rules
  account: remove ASCII whitespace at both ends; an empty result is missing.
  Quarantine each invalid payment with the FIRST failing reason:
    1. missing_account   account is missing
    2. unknown_account   account has no matching row in accounts
  Valid payments get the account's name.

Outputs (row order does not matter)
  clean(id text, account text, name text, amount_cents int64)
  exceptions(id text, reason text)