Oracle SQL仅返回具有不同列值的行

Question

I have two tables table a and table b with exact same structure. I would like to return the columns with different values: Could somebody tell me how to achieve the below

for eg;

Table A

ID, Fname, Lname, Age 
1   Mike   Smith  10

Table B

ID, Fname, Lname, Age  
1   Mike   Smith  20

My result should be

different_columns

Age 

Answer 1

select id,
       listagg(diff_col, ', ') within group(order by diff_col) as diff_cols
  from (select a.id, 'fname' as diff_col
          from tablea a
          left join tableb b
            on a.id = b.id
           and a.fname = b.fname
         where b.id is null
        union all
        select a.id, 'lname'
          from tablea a
          left join tableb b
            on a.id = b.id
           and a.lname = b.lname
         where b.id is null
        union all
        select a.id, 'age'
          from tablea a
          left join tableb b
            on a.id = b.id
           and a.age = b.age
         where b.id is null)
 group by id

Fiddle: http://sqlfiddle.com/#!4/daeaf/4/0

(I added an ID # 2 with a different age AND lname, for illustration)

Another way of doing this is a giant case statement but I would prefer the above as it's easier to tack on more conditions should your actual table have more than 3 columns with which you are doing this test.

If you don't mind having one column per column check (see next fiddle) this might be more ideal for checking that many columns. I'm not sure there is an easier way; you'd have to tack on a bunch of ORs to the WHERE clause and a bunch more case statements for each column comparison.

select a.id,
       case when a.fname <> b.fname then 'X' end as fname,
       case when a.lname <> b.lname then 'X' end as lname,
       case when a.age <> b.age then 'X' end as age
  from tablea a
  join tableb b
    on a.id = b.id
 where a.fname <> b.fname
    or a.lname <> b.lname
    or a.age <> b.age

Fiddle: http://sqlfiddle.com/#!4/daeaf/7/0

(this shows an X if there is a difference in the indicated column, otherwise nothing)

Answer 2

If you union both tables and group by id , then you'll know that any group having count(*) > 1 has at least 1 column that's different (since union automatically removes duplicates).

In order to determine which column is different, you can use count(distinct column_name) to see which columns have more than 1 distinct value.

select id,
    count(distinct fname) fname,
    count(distinct lname) lname,
    count(distinct age) age
from (
    select * from tablea
    union
    select * from tableb
) t1 group by id
having count(*) > 1