Pfthb

I'm looking to only return rows where if my first condition doesn't exist
then return new condition from same table, and additional conditions:
basically...:

SELECT * 
FROM Table1 
WHERE

Condition#1: Column1=Column2

Condition#2: AND IF Column1 = Column2 doesn't exist (matching values for same ID) 
 THEN return results where Column3 = Column4

Thanks.

First you need to check what case you are. So you count how many C1 = C2 you have in your table

SELECT [id], COUNT(CASE WHEN C1 = C2 THEN 1 END) as c1_c2
FROM Table1
GROUP BY [id]

Then you have two conditions

first when you have some c1 = c2 you do

 (c1 = c2 and c1_c2 > 0 )

and the second when dont have any c1 = c2

 (c1_c2 = 0 and c3 = c4)

SQL DEMO:: Because is an OR and the c1_c2 count decide which side of the where is going to work.

WITH c12 as (
 SELECT [id], COUNT(CASE WHEN C1 = C2 THEN 1 END) as c1_c2_cnt
 FROM Table1
 GROUP BY [id]
) 
SELECT *
FROM Table1 T
JOIN c12 C
 ON T.[id] = C.[id] 
WHERE ( C.c1_c2_cnt > 0 and T.c1 = T.c2 )
 OR ( C.c1_c2_cnt = 0 and T.c3 = T.c4 )

OUTPUT

I suspect that this will work for what the OP is after, however, the logic seems odd...

SELECT Columns
FROM YourTable YT
WHERE YT.Column1 = YT.Column2
 OR (YT.Column3 = YT.Column4
 AND NOT EXISTS(SELECT 1
 FROM YourTable e
 WHERE YT.id = e.id
 AND e.Column1 = e.Column2));

Not sure how good the performance will be here though, as you may well end up with a table scan with that NOT EXISTS.

Well, based off the comments, here is one way.

SELECT * 
FROM Table1 
WHERE Column1 = Column2 
UNION 
SELECT * 
FROM Table2 
WHERE Column1 != Column2 and Column3 = Column4

You can do this by using simple booleans:

WHERE
Column1 = Column2
OR
Column3 = Column4

The Venn diagram for this:

All of the shaded area will be returned, including when Column1 = Column2 & Column3 = Column4 (the dark shaded intersection)

If you don't want rows where both conditions are met then you can modify the logic:

WHERE
Column1 = Column2
OR
(Column1 <> Column2 
AND
Column3 = Column4)

The Venn would look the same except the intersection would be white and excluded.

Another approach :

DECLARE 
 @t TABLE( ID INT, col1 CHAR(1), col2 CHAR(1), col3 INT, col4 INT)

INSERT INTO @t VALUES 
(1234,'A','B',100,100),
(1234,'A','A',100,100),
(2468,'A','C',125,125),
(2468,'A','B',100,150)

SELECT 
 ID
, col1
, col2
, col3
, col4
FROM (
 SELECT *, ROW_NUMBER() OVER(PARTITION BY ID, col1 ORDER BY col2) RN
 FROM (
 SELECT *
 FROM @t
 WHERE 
 col3 = col4
 AND col1 = col2
 UNION ALL 
 SELECT *
 FROM @t
 WHERE 
 col3 = col4
 AND ID NOT IN(
 SELECT ID
 FROM @t
 WHERE 
 col3 = col4
 AND col1 = col2
 )
 ) D
) E
WHERE 
 RN = 1