If you look at this SQL article [old link removed], you may either be impressed or insult me for bad coding practise. FYI, it does run on 1000 rows in < 1 second. When I was using MsSql (before I migrated to CentOS/Mono/MySql) I sometimes had to go to extortionate amounts of effort to get stuff done, but with MySql it all seems, a bit too easy.. So from the giant SQL that you saw in the previous article, I can do all that on the same tables (bar one change) with the following SQL:
SELECT t_Hosts.Id AS Id, t_Records.Id AS RecordId, Hostname,
Frequency, PingStatus, PingLatency, HttpStatusCode,
HttpOnlineStatus, HttpLatency, CheckDate
FROM t_Hosts
LEFT JOIN t_Records ON t_Hosts.Id = t_Records.HostId
WHERE t_Records.CheckDate IS NULL
OR t_Records.CheckDate < DATE_SUB(NOW(), INTERVAL t_Hosts.Frequency MINUTE);
This happens to have caused more trouble than it's worth, so it's back to a datetime now.
Also, make sure that you add an alias for the second Id or it may overwrite the first tables Id!
Edit: It would seem this doesn't work, and is actually flawed. My old logic (@mcfly version) is actually the way to go, not streakys stupid left join crap!
Below is the MySQL Version.
SELECT t_Hosts.Id AS Id, MAX(Hostname) AS Hostname, MAX(Frequency) AS Frequency
FROM t_Hosts
LEFT JOIN t_Records
ON t_Hosts.Id = t_Records.HostId
GROUP BY t_Hosts.Id
HAVING MAX(CheckDate) <= DATE_SUB(NOW(), INTERVAL MAX(Frequency) MINUTE)
UNION
SELECT Id, Hostname, Frequency
FROM t_Hosts
WHERE (
Id NOT IN (
SELECT HostId
FROM t_Records
WHERE (HostId IS NOT NULL)
)
)
Top comments (0)