Question Re: Table data comparison in Technology Q&A

Table data comparison

Former Member — Mon, 18 Apr 2011 09:43:33 GMT

When comparing data (fields) between two tables is there an advantage to using EXCEPT, INTERSECT, or, in the case of looking for data in ONE field between two tables, can a JOIN of the two tables accomplish the same thing?

ex: //join two tables and look for matching fields: select t.FIELD, t2.FIELD from tableone t join tabletwo t where t.FIELD = t2.FIELD

ex2: //two tables comparing the same two fields using intersect:

ex3: //two tables comparing the same two fields using except:

Appreciate any help provided.

Re: Table data comparison

Former Member — Mon, 18 Apr 2011 10:14:10 GMT

The answer will depend on what the question is - are you looking for rows that are the same, rows that are different, or both?

See the answers in this thread that can get you started.

Re: Table data comparison

Former Member — Mon, 18 Apr 2011 11:07:09 GMT

I'll try to clarify. If you are looking in two tables to see if data in one field is the same in both tables is there a benefit to using a join or INTERSECT/EXCEPT. This could be if the fields are the same, different or both. Generally speaking. Will a join provide the same results as INTERSECT/EXCEPT?

Re: Table data comparison

Former Member — Mon, 18 Apr 2011 11:29:54 GMT

A join, by definition, will produce a result set with matches. EXCEPT will produce a result set containing differences.

INTERSECT is roughly equivalent to inner join except that:

INTERSECT handles NULL values quite differently than join does, and
the number of rows you get in the result won't match the join either, in general.

Re: Table data comparison

Former Member — Mon, 18 Apr 2011 11:34:45 GMT

I think I got it. Difference being the handling of NULL. Much appreciated.

Re: Table data comparison

VolkerBarth — Mon, 18 Apr 2011 11:49:58 GMT

That being said, an OUTER JOIN can also be used to produce a difference, as long as you just select those rows that have no match (or differ in other ways), say

select T1.pk, T2.pk from T1 full outer join T2 on T1.pk = T2.pk where T1.pk is null or T2.pk is null

(However, IMHO, EXCEPT would be much more comprehensible here.)