I have a database with a table called QuizMatches. The table has the following structure:
CREATE TABLE `QuizMatches` (
`QuizMatchesGuid` binary(16) NOT NULL,
`DateStarted` datetime NOT NULL,
`LatestChanged` datetime NOT NULL,
`HostFBUserToken` varchar(250) NOT NULL,
`GuestFBUserToken` varchar(250) NOT NULL,
`ArrayOfQuestionIDs` varchar(200) NOT NULL,
`ArrayOfQuestionResponseTimesAndAnswersHost` varchar(900) NOT NULL,
`ArrayOfQuestionResponseTimesAndAnswersGuest` varchar(900) NOT NULL,
`MatchFinished` int(1) NOT NULL DEFAULT '0',
`Category` varchar(45) NOT NULL,
`JsonQuestions` varchar(4000) NOT NULL DEFAULT '[]',
`DateFinished` datetime NOT NULL,
`LatestPushSentDate` datetime NOT NULL DEFAULT '0000-00-00 00:00:00',
PRIMARY KEY (`QuizMatchesGuid`),
KEY `HostFBUserTokenIX` (`HostFBUserToken`),
KEY `GuestFBUserTokenIX` (`GuestFBUserToken`),
KEY `MatchFinishedIX` (`MatchFinished`),
KEY `LatestChangedIX` (`LatestChanged`),
KEY `LatestPushSentDateIX` (`LatestPushSentDate`),
KEY `DateFinishedIX` (`LatestChanged`,`HostFBUserToken`,`GuestFBUserToken`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8
There is a large number of rows in this table and it is heavily used by multiple clients especially queries like the following are executed:
SELECT HEX(QuizMatchesGuid) AS QuizMatchesGuid, DateStarted,
LatestChanged, HostFBUserToken, GuestFBUserToken,
ArrayOfQuestionIDs, ArrayOfQuestionResponseTimesAndAnswersHost,
ArrayOfQuestionResponseTimesAndAnswersGuest, JsonQuestions
FROM CrystalDBQuiz.QuizMatches
ORDER BY LatestChanged DESC
LIMIT 10
The main problem seem to be that the database performs a full index scan. I have tried with different combinations of indexes but to no success.
If I run EXPLAIN on the above SELECT query I receive the following:
id: 1
select_type: SIMPLE
table: 'QuizMatches'
type: index
possible_keys: NULL
key: 'LatestChangedIX'
key_len: 8
ref: NULL
rows: 10
Extra:
Is there a way I can optimize SELECTS as the above example towards this database table?
If you are using the
LIMIT
statement for pagination I suggest you to useLatestChanged
value for this due to ordering. So your query will turn to