select *
from (
select 'T1' T, *
from DB1.dbo.Table
except
select 'T2' T, *
from DB2.dbo.Table
) as T
union all
select *
from (
select 'T2' T, *
from DB2.dbo.Table
except
select 'T1' T, *
from DB1.dbo.Table
) as T
ORDER BY 2,3,4, ..., 1 -- make T1 and T2 to be close in output 2,3,4 are UNIQUE KEY SEGMENTS
Test code:
declare @T1 table (ID int)
declare @T2 table (ID int)
insert into @T1 values(1),(2)
insert into @T2 values(2),(3)
select *
from (
select *
from @T1
except
select *
from @T2
) as T
union all
select *
from (
select *
from @T2
except
select *
from @T1
) as T
Result:
ID
-----------
1
3
Note: It can take long time to compare big table, when developing “tuned” solution or refactorig, which will give same result as REFERERCE – it may be wise to chekc simple parameters first: like
select count(t.*) from (
select count(*) c0, SUM(BINARY_CHECKSUM(*)%1000000) c1 FROM T_REF_TABLE
-- select 12345 c0, -214365454 c1 -- constant values FROM T_REF_TABLE
except
select count(*) , SUM(BINARY_CHECKSUM(*)%1000000) FROM T_WORK_COPY
) t
When this is empty, you have probably things under controll, and may be you can modify when you fail you will see “constant values FROM T_REF” to isert to save even more time for next check!!!