compare two tables's data in MySQL [duplicate] - mysql

Question

This question already has answers here:
Closed 10 years ago.
There are two tables
Table 1
select * from manage_tcp;
+----+----------+--------+
| id | services | statut |
+----+----------+--------+
| 2 | 1433 | up |
| 2 | 3389 | up |
| 3 | 3389 | down |
| 3 | 1433 | down |
| 9 | 3389 | up |
| 8 | 3389 | up |
| 7 | 1433 | up |
| 6 | 3389 | up |
| 5 | 3389 | up |
| 4 | 3389 | up |
| 10 | 1433 | up |
| 11 | 1433 | up |
| 12 | 3389 | up |
| 13 | 1433 | up |
| 14 | 3389 | up |
| 15 | 1433 | up |
| 16 | 3389 | up |
| 17 | 1433 | up |
| 18 | 3389 | up |
| 19 | 1433 | up |
| 20 | 3389 | up |
| 21 | 1433 | up |
| 24 | 1433 | up |
| 23 | 3389 | up |
| 25 | 3389 | up |
| 26 | 1433 | up |
| 29 | 1433 | up |
| 28 | 3389 | up |
| 30 | 3389 | up |
| 31 | 80 | up |
| 32 | 3389 | up |
| 33 | 80 | up |
| 34 | 3389 | up |
| 35 | 80 | up |
+----+----------+--------+
34 rows in set (0.00 sec)
Table 2
mysql> select * from manage_host;
+----+------------+-------------+--------+-------+--------------+----------+------+
| id | uptime | type | statut | group | thresold_ref | thresold | mail |
+----+------------+-------------+--------+-------+--------------+----------+------+
| 2 | 277248529 | win2003.png | up | 4 | 1 | 0 | NULL |
| 3 | 277277471 | win2003.png | prob | 4 | 1 | 0 | NULL |
| 4 | 346159833 | win2003.png | up | 4 | 1 | 0 | NULL |
| 5 | 930205491 | win2003.png | up | 5 | 1 | 0 | NULL |
| 6 | 3805663007 | win2003.png | up | 5 | 1 | 0 | NULL |
| 7 | -1 | win2003.png | up | 4 | 1 | 0 | NULL |
| 8 | 29413867 | win2003.png | up | 5 | 1 | 0 | NULL |
| 9 | 981986401 | win2003.png | up | 5 | 1 | 0 | NULL |
| 10 | -1 | win2003.png | up | 5 | 1 | 0 | NULL |
| 11 | -1 | win2003.png | up | 5 | 1 | 0 | NULL |
| 12 | 2230787611 | win2003.png | up | 7 | 1 | 0 | NULL |
| 13 | -1 | win2003.png | up | 7 | 1 | 0 | NULL |
| 14 | 1004161923 | win2003.png | up | 8 | 1 | 0 | NULL |
| 15 | -1 | win2003.png | up | 8 | 1 | 0 | NULL |
| 16 | 1592294954 | win2003.png | up | 8 | 1 | 0 | NULL |
| 17 | -1 | win2003.png | up | 8 | 1 | 0 | NULL |
| 18 | 1449216250 | win2003.png | up | 4 | 1 | 0 | NULL |
| 19 | -1 | win2003.png | up | 4 | 1 | 0 | NULL |
| 20 | 3461945234 | win2003.png | up | 4 | 1 | 0 | NULL |
| 21 | -1 | win2003.png | up | 4 | 1 | 0 | NULL |
| 23 | 3461946968 | win2003.png | up | 4 | 1 | 0 | NULL |
| 24 | -1 | win2003.png | up | 4 | 1 | 0 | NULL |
| 25 | 3462022562 | win2003.png | up | 4 | 1 | 0 | NULL |
| 26 | -1 | win2003.png | up | 4 | 1 | 0 | NULL |
| 35 | 3318333401 | win2003.png | up | 6 | 1 | 0 | |
| 28 | 3461879984 | win2003.png | up | 4 | 1 | 0 | NULL |
| 29 | -1 | win2003.png | up | 4 | 1 | 0 | NULL |
| 30 | 1872950109 | win2003.png | up | 6 | 1 | 0 | NULL |
| 31 | 1872950125 | win2003.png | up | 6 | 1 | 0 | NULL |
| 32 | 1955232033 | win2003.png | up | 6 | 1 | 0 | NULL |
| 34 | 3318333401 | win2003.png | up | 6 | 1 | 0 | |
| 33 | 1955232049 | win2003.png | up | 6 | 1 | 0 | NULL |
+----+------------+-------------+--------+-------+--------------+----------+------+
32 rows in set (0.00 sec)
I'd like to compare those two tables above and to get the results in table 1 but not in table 2 like: (this is the result i want)
+----+----------+--------+
| id | services | statut |
+----+----------+--------+
| 2 | 1433 | up |
| 3 | 3389 | down |
| 3 | 1433 | down |
+----+----------+--------+
I've tried
SELECT a.*,b.*
FROM manage_tcp as a
LEFT JOIN manage_host as b
USING(id,statut)
WHERE b.id is null;
but I got
+----+----------+--------+
| id | services | statut |
+----+----------+--------+
| 3 | 3389 | down |
| 3 | 1433 | down |
+----+----------+--------+
it's not what I want.
Thanks a lot!

Answer 1

Try this:
SELECT m.id, m.services, m.statut
FROM manage_tcp m
WHERE EXISTS
(
SELECT *
FROM manage_host mh
WHERE mh.id = m.id AND m.statut = mh.statut
)
This will give you the intersection of the tables manage_tcp and manage_host based on id and status columns.
get the results in table 1 but not in table 2
If you need the difference between the tow tables, use NOT EXISTS instead of EXISTS like this:
SELECT m.id, m.services, m.statut
FROM manage_tcp m
WHERE NOT EXISTS
(
SELECT *
FROM manage_host mh
WHERE mh.id = m.id AND m.statut = mh.statut
)
Typically, this is the implementations of the INTERSECT and EXCEPT operators which are defined by the SQL standard as the set intersection and difference operators. But, unfortunately, most database system doesn't implement these operators. MS SQL Server does support these operators but not MYSQL .

Answer 2

Try this: (assume the comparison column are id and statut)
SELECT a.id, a.services, a.statut
FROM manage_tcp a
WHERE NOT EXISTS(SELECT b.id FROM manage_host b WHERE b.id=a.id AND b.statut=a.statut)
UNION
SELECT a.id, (SELECT services FROM manage_tcp WHERE id=a.id ORDER BY id LIMIT 1), a.statut
FROM manage_tcp a
GROUP BY a.id, a.statut HAVING COUNT(1)>1
The first query will produce result from table a where not in table b.
The second, after UNION, will produce result for id that appear twice in table a.

Answer 3

SELECT manage_tcp.* FROM
(
SELECT a.id,a.statut,a.cnt
FROM (
SELECT id, statut, COUNT(1) AS cnt FROM manage_tcp GROUP BY id,statut
) AS a
LEFT JOIN manage_host AS b USING(id, statut)
GROUP BY a.id, a.statut
HAVING a.cnt > COUNT(1)
) AS diff
LEFT JOIN manage_tcp USING(id,statut)
or
SELECT a.id,a.statut,a.services,a.cnt
FROM (
SELECT id,statut,services,COUNT(1) AS cnt FROM manage_tcp GROUP BY id,statut
) AS a
LEFT JOIN manage_host AS b USING(id, statut)
GROUP BY a.id,a.statut
HAVING a.cnt > COUNT(1)