将 3 个案例陈述合并或简化为一个

Question

我会将以下 3 个 case 语句合并为一个并尽可能简化。查询的目的是根据 3 列找出任何具有无效或有效电话号码的成员。如果任何列具有有效的电话号码，则它是有效的。如果全部无效，则视为无效。任何帮助都会很棒。

    CASE 
        WHEN REGEXP_REPLACE(DAY_PH, '[^[:digit:]'''']', NULL) IS NULL
            THEN 1
  WHEN REGEXP_REPLACE(DAY_PH, '[^[:digit:]'''']', NULL) ='9999999999'
    THEN 1
        WHEN LENGTH(TRIM(REGEXP_REPLACE(DAY_PH, '[^[:digit:]'''']', NULL))) = 10
            THEN 0
        ELSE 1
        END INVALID_DAY_PH,
    CASE 
        WHEN REGEXP_REPLACE(EVE_PH, '[^[:digit:]'''']', NULL) IS NULL
            THEN 1
  WHEN  REGEXP_REPLACE(EVE_PH, '[^[:digit:]'''']', NULL)='9999999999'
    THEN 1  
        WHEN LENGTH(TRIM(REGEXP_REPLACE(EVE_PH, '[^[:digit:]'''']', NULL))) = 10
            THEN 0
        ELSE 1
        END INVALID_EVE_PH,
    CASE 
        WHEN REGEXP_REPLACE(MOB_PH, '[^[:digit:]'''']', NULL) IS NULL
            THEN 1
   WHEN REGEXP_REPLACE(MOB_PH, '[^[:digit:]'''']', NULL)='9999999999'
    THEN 1  
        WHEN LENGTH(TRIM(REGEXP_REPLACE(MOB_PH, '[^[:digit:]'''']', NULL))) = 10
            THEN 0
        ELSE 1
        END INVALID_MOB_PH,

Answer 1

您可以通过将正则表达式调用推入内联视图或 CTE 来简化一点，这样您就不必重复它们：

select id, day_ph, eve_ph, mob_ph,
  case when length(day_ph_cln) = 10 and day_ph_cln != '9999999999'
    then 0 else 1 end as invalid_day_ph,
  case when length(eve_ph_cln) = 10 and eve_ph_cln != '9999999999'
    then 0 else 1 end as invalid_eve_ph,
  case when length(mob_ph_cln) = 10 and mob_ph_cln != '9999999999'
    then 0 else 1 end as invalid_mob_ph
from (
  select id, day_ph, eve_ph, mob_ph,
    regexp_replace(day_ph, '[^[:digit:]'''']') as day_ph_cln,
    regexp_replace(eve_ph, '[^[:digit:]'''']') as eve_ph_cln,
    regexp_replace(mob_ph, '[^[:digit:]'''']') as mob_ph_cln
  from your_table
);

trim() 似乎没有做任何事情，所以我把它省略了 - 正则表达式已经删除了所有空格，因为它们不是数字（或单引号......我已经离开了模式完好无损，但不确定它是否有意义......）

如果您想要一个最终标志来显示是否所有其他标志都无效，那么您可以使用least()，如果其他三个标志中的任何一个为零，它将给您零，否则为 1。

select id, day_ph, eve_ph, mob_ph,
  case when length(day_ph_cln) = 10 and day_ph_cln != '9999999999'
    then 0 else 1 end as invalid_day_ph,
  case when length(eve_ph_cln) = 10 and eve_ph_cln != '9999999999'
    then 0 else 1 end as invalid_eve_ph,
  case when length(mob_ph_cln) = 10 and mob_ph_cln != '9999999999'
    then 0 else 1 end as invalid_mob_ph,
  least(
    case when length(day_ph_cln) = 10 and day_ph_cln != '9999999999'
      then 0 else 1 end,
    case when length(eve_ph_cln) = 10 and eve_ph_cln != '9999999999'
      then 0 else 1 end,
    case when length(mob_ph_cln) = 10 and mob_ph_cln != '9999999999'
      then 0 else 1 end
  ) as all_invalid
from (
  select id, day_ph, eve_ph, mob_ph,
    regexp_replace(day_ph, '[^[:digit:]'''']') as day_ph_cln,
    regexp_replace(eve_ph, '[^[:digit:]'''']') as eve_ph_cln,
    regexp_replace(mob_ph, '[^[:digit:]'''']') as mob_ph_cln
  from your_table
);

或使用其他级别的子查询（或 CTE）来删除重复：

select id, day_ph, eve_ph, mob_ph,
  invalid_day_ph, invalid_eve_ph, invalid_mob_ph,
  least(invalid_day_ph, invalid_eve_ph, invalid_mob_ph) as all_invalid
from (
  select id, day_ph, eve_ph, mob_ph,
    case when length(day_ph_cln) = 10 and day_ph_cln != '9999999999'
      then 0 else 1 end as invalid_day_ph,
    case when length(eve_ph_cln) = 10 and eve_ph_cln != '9999999999'
      then 0 else 1 end as invalid_eve_ph,
    case when length(mob_ph_cln) = 10 and mob_ph_cln != '9999999999'
      then 0 else 1 end as invalid_mob_ph
  from (
    select id, day_ph, eve_ph, mob_ph,
      regexp_replace(day_ph, '[^[:digit:]'''']') as day_ph_cln,
      regexp_replace(eve_ph, '[^[:digit:]'''']') as eve_ph_cln,
      regexp_replace(mob_ph, '[^[:digit:]'''']') as mob_ph_cln
    from your_table
  )
);

如果您只想要最后一个标志，那么您可以组合大小写表达式：

select id, day_ph, eve_ph, mob_ph,
  case when (length(day_ph_cln) = 10 and day_ph_cln != '9999999999')
         or (length(eve_ph_cln) = 10 and eve_ph_cln != '9999999999')
         or (length(mob_ph_cln) = 10 and mob_ph_cln != '9999999999')
    then 0 else 1 end as all_invalid
from (
  select id, day_ph, eve_ph, mob_ph,
    regexp_replace(day_ph, '[^[:digit:]'''']') as day_ph_cln,
    regexp_replace(eve_ph, '[^[:digit:]'''']') as eve_ph_cln,
    regexp_replace(mob_ph, '[^[:digit:]'''']') as mob_ph_cln
  from your_table
);

您没有提供任何样本数据来验证这是否提供了您想要的结果；但是here's a db<>fiddle 有一些非常基本的虚构数据。