| 枚举 | 状态、渠道、类型字段的取值集合 |
|---|---|
| 范围 | 金额非负、比例在 0 到 1 之间 |
| 格式 | 手机号、日期字符串、编码位数 |
| 处置 | 字典外值告警并按策略转 unknown |
取值分布检查
取值分布检查
-- 列出各状态取值与占比,出现字典外取值即异常
SELECT status_name, COUNT(1) AS cnt
FROM dwd_trade_order_detail_di
WHERE dt = '2026-10-11'
GROUP BY status_name
ORDER BY cnt DESC;范围与格式
范围与格式
SELECT COUNT(1) AS bad_rows
FROM dwd_trade_order_detail_di
WHERE dt = '2026-10-11'
AND (pay_amount < 0 OR (phone IS NOT NULL AND phone NOT RLIKE '^1[3-9][0-9]{9}$'));字典外取值不要静默丢弃:先落档并计数,再决定清洗规则,否则问题会被隐藏。