目录

需求

实现

mysql

oracle

pg

hive

达梦

---

需求

 在数据质量检核中，经常会要求检查数据数据符合某些要求，比较某个字段是否是数字或者是否是日期等等，经常这种就用到数据库正则

实现

参数说明：

通用规则，${PARAM}换成合适的正则表达式，${TABLE}换成实际数据表，${COL}换成实际字段

mysql

校验手机号

select (b - a) as c0, b as c1,(case when b=0 then 0 else round((b-a)*1.0/b,2) end) as c2 from (( select count(1) as a from ${TABLE} where  ${COL}  regexp '^((13[0-9])|(14(0|[5-7]|9))|(15([0-3]|[5-9]))|(16(2|[5-7]))|(17[0-8])|(18[0-9])|(19([0-3]|[5-9])))[0-9]{8}$' ) a, (select count(1) as b from ${TABLE}) b)

oracle

校验手机号

select (b - a) as c0, b as c1,(case when b=0 then 0 else round((b-a)*1.0/b,2) end) as c2 from ( select count(1) as a from ${TABLE} where regexp_like(${COL},'^((13[0-9])|(14(0|[5-7]|9))|(15([0-3]|[5-9]))|(16(2|[5-7]))|(17[0-8])|(18[0-9])|(19([0-3]|[5-9])))[0-9]{8}$')) a, (select count(1) as b from ${TABLE}) b

pg

select sum(case when cast(${COL} as varchar)  ~${PARAM} then 0 else 1 end)  as c0,count(*) as c1  from ${TABLE} where ${COL} is not null

hive

通用规则，${PARAM}换成合适的正则表达式，${TABLE}换成实际数据表，${COL}换成实际字段

select sum(case when cast(${COL} as varchar) regexp ${PARAM} then 0 else 1 end)  as c0,count(*) as c1  from ${TABLE}

达梦

校验手机号

select nvl(b - a,0),nvl(a,0), nvl(b,0), decode(b,0,0,(b-a)*1.00/b) from ( select count(1) as a from ${TABLE} where REGEXP_LIKE(${COL}, '^((13[0-9])|(14[5|7])|(15([0-3]|[5-9]))|(17[013678])|(18[0,5-9]))\d{8}$') ) a,( select count(1) as b from ${TABLE}) b