Do you wonder if your IT business is truly profitable or if you should raise your prices? Learn how to calculate your overhead burden using our free interactive tool and use it to determine the right price for your IT services. Start calculating Now!
How to get distinct list when all records are true
I have a table as below
Select Id, Feld1, Feld2, Feld3 from table1
How can I get distinct list of Ids when either ( All Feld1 = 'Y' or All Feld2 = 'Y' or All Feld2 = 'Y' )
Id, Feld1, Feld2, Feld3
1 Y, N, N
1 N, N, N
2 Y, N, Y
2 Y, N, N
3 N, Y, N
3 N, Y, N
3 N, Y, N
Should return 2, 3
Id, Feld1, Feld2, Feld3
1 Y, N, N
2 Y, N, Y
2 Y, N, N
3 N, Y, N
3 N, Y, N
3 N, Y, N
Should return 1, 2, 3
Select Id, Feld1, Feld2, Feld3 from table1
How can I get distinct list of Ids when either ( All Feld1 = 'Y' or All Feld2 = 'Y' or All Feld2 = 'Y' )
Id, Feld1, Feld2, Feld3
1 Y, N, N
1 N, N, N
2 Y, N, Y
2 Y, N, N
3 N, Y, N
3 N, Y, N
3 N, Y, N
Should return 2, 3
Id, Feld1, Feld2, Feld3
1 Y, N, N
2 Y, N, Y
2 Y, N, N
3 N, Y, N
3 N, Y, N
3 N, Y, N
Should return 1, 2, 3
Do more with
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
Does this not work?
SELECT DISTINCT Id, Field1, Field2, Field3
FROM YourTable
WHERE Field1 = 'Y' OR Field2='Y' OR FIeld3='Y'
or would it be something like this:
DECLARE @data TABLE (
id INT,
field1 CHAR(1),
field2 CHAR(1),
field3 CHAR(1)
)
INSERT INTO @data ( id, field1, field2, field3 )
VALUES
(1, 'Y', 'N', 'N' ),
(1, 'N', 'N', 'N' ),
(2, 'Y', 'N', 'Y' ),
(2, 'Y', 'N', 'N' ),
(3, 'N', 'Y', 'N' ),
(3, 'N', 'Y', 'N' ),
(3, 'N', 'Y', 'N' )
--SELECT * FROM @data
; WITH cte
AS (
SELECT A.id, COUNT(*) AS NbRows
, (SELECT COUNT(*) FROM @data WHERE id = A.id AND field1 = 'Y') AS F1
, (SELECT COUNT(*) FROM @data WHERE id = A.id AND field2 = 'Y') AS F2
, (SELECT COUNT(*) FROM @data WHERE id = A.id AND field3 = 'Y') AS F3
FROM @data AS A
GROUP BY A.id
)
SELECT cte.id
FROM cte
WHERE cte.NbRows = cte.F1 OR cte.NbRows = cte.F2 OR cte.NbRows = cte.F3
DELETE FROM @data
INSERT INTO @data ( id, field1, field2, field3 )
VALUES
(1, 'Y', 'N', 'N' ),
(2, 'Y', 'N', 'Y' ),
(2, 'Y', 'N', 'N' ),
(3, 'N', 'Y', 'N' ),
(3, 'N', 'Y', 'N' ),
(3, 'N', 'Y', 'N' )
--SELECT * FROM @data
; WITH cte
AS (
SELECT A.id, COUNT(*) AS NbRows
, (SELECT COUNT(*) FROM @data WHERE id = A.id AND field1 = 'Y') AS F1
, (SELECT COUNT(*) FROM @data WHERE id = A.id AND field2 = 'Y') AS F2
, (SELECT COUNT(*) FROM @data WHERE id = A.id AND field3 = 'Y') AS F3
FROM @data AS A
GROUP BY A.id
)
SELECT cte.id
FROM cte
WHERE cte.NbRows = cte.F1 OR cte.NbRows = cte.F2 OR cte.NbRows = cte.F3
SELECT id
FROM @data
GROUP BY id
HAVING
MIN(CASE WHEN field1 = 'Y' THEN 1 ELSE 0 END) = 1 OR
MIN(CASE WHEN field2 = 'Y' THEN 1 ELSE 0 END) = 1 OR
MIN(CASE WHEN field3 = 'Y' THEN 1 ELSE 0 END) = 1
FROM @data
GROUP BY id
HAVING
MIN(CASE WHEN field1 = 'Y' THEN 1 ELSE 0 END) = 1 OR
MIN(CASE WHEN field2 = 'Y' THEN 1 ELSE 0 END) = 1 OR
MIN(CASE WHEN field3 = 'Y' THEN 1 ELSE 0 END) = 1
I believe Scott's approach will efficiently solve the question as asked.
I just wanted to show an alternative that would also allow determination of which column has all 'Y's
It starts by "unpivoting" the data (but not by using the "unpivot" command, I find the following to be simpler syntax for that). Once that is complete it sis is just a group by query to complete the task.
I just wanted to show an alternative that would also allow determination of which column has all 'Y's
It starts by "unpivoting" the data (but not by using the "unpivot" command, I find the following to be simpler syntax for that). Once that is complete it sis is just a group by query to complete the task.
CREATE TABLE mytable(
Id INTEGER NOT NULL
,Field1 VARCHAR(4) NOT NULL
,Field2 VARCHAR(2) NOT NULL
,Field3 VARCHAR(2) NOT NULL
);
INSERT INTO mytable(Id,Field1,Field2,Field3) VALUES (1,'Y','N','N');
INSERT INTO mytable(Id,Field1,Field2,Field3) VALUES (1,'N','N','N');
INSERT INTO mytable(Id,Field1,Field2,Field3) VALUES (2,'Y','N','Y');
INSERT INTO mytable(Id,Field1,Field2,Field3) VALUES (2,'Y','N','N');
INSERT INTO mytable(Id,Field1,Field2,Field3) VALUES (3,'N','Y','N');
INSERT INTO mytable(Id,Field1,Field2,Field3) VALUES (3,'N','Y','N');
INSERT INTO mytable(Id,Field1,Field2,Field3) VALUES (3,'N','Y','N');
select ca.id, ca.all_y
from mytable
cross apply (
values
(id, 'field1', field1)
, (id, 'field2', field2)
, (id, 'field3', field3)
) ca (id, all_y, fieldval)
group by ca.id, ca.all_y
having min(fieldval) = max(fieldval) and max(fieldval) = 'Y'
;
+-----+--------+
| id | all_y |
+-----+--------+
| 2 | field1 |
| 3 | field2 |
+-----+--------+
hello i would like to recommend a differend possibility.
in case of most records are valid, it's a good idea to search for records which are not matching. In addition then you have a cobination of AND which is faster then OR.
so may be:
SELECT DISTINCT ID from mytable where ID not in (
SELECT ID from mytable where field1 = 'N' AND field2 = 'N' AND field3 = 'N'
)
please check finaly execution plan to see what is your optimal solution.
Good luck
in case of most records are valid, it's a good idea to search for records which are not matching. In addition then you have a cobination of AND which is faster then OR.
so may be:
SELECT DISTINCT ID from mytable where ID not in (
SELECT ID from mytable where field1 = 'N' AND field2 = 'N' AND field3 = 'N'
)
please check finaly execution plan to see what is your optimal solution.
Good luck
If you want to know which column matched, you can adjust my query as follows:
SELECT id,
CASE WHEN MIN(CASE WHEN field1 = 'Y' THEN 10 ELSE 0 END) = 10 THEN 'Field1'
WHEN MIN(CASE WHEN field2 = 'Y' THEN 20 ELSE 0 END) = 20 THEN 'Field2'
WHEN MIN(CASE WHEN field3 = 'Y' THEN 30 ELSE 0 END) = 30 THEN 'Field3'
END AS which_field_matched
FROM @data
GROUP BY id
HAVING
MIN(CASE WHEN field1 = 'Y' THEN 10 ELSE 0 END) = 10 OR
MIN(CASE WHEN field2 = 'Y' THEN 20 ELSE 0 END) = 20 OR
MIN(CASE WHEN field3 = 'Y' THEN 30 ELSE 0 END) = 30
The CROSS APPLY is also fine.
FYI, the "NOT IN" method will require two full scans of the table; I designed my method to insure that it requires only one table scan, and the CROSS APPLY should require only one table scan also.
SELECT id,
CASE WHEN MIN(CASE WHEN field1 = 'Y' THEN 10 ELSE 0 END) = 10 THEN 'Field1'
WHEN MIN(CASE WHEN field2 = 'Y' THEN 20 ELSE 0 END) = 20 THEN 'Field2'
WHEN MIN(CASE WHEN field3 = 'Y' THEN 30 ELSE 0 END) = 30 THEN 'Field3'
END AS which_field_matched
FROM @data
GROUP BY id
HAVING
MIN(CASE WHEN field1 = 'Y' THEN 10 ELSE 0 END) = 10 OR
MIN(CASE WHEN field2 = 'Y' THEN 20 ELSE 0 END) = 20 OR
MIN(CASE WHEN field3 = 'Y' THEN 30 ELSE 0 END) = 30
The CROSS APPLY is also fine.
FYI, the "NOT IN" method will require two full scans of the table; I designed my method to insure that it requires only one table scan, and the CROSS APPLY should require only one table scan also.
Premium Content
You need an Expert Office subscription to comment.Start Free Trial