sql query is slow - php

Question

I have a phpmyadmin database with 1 000 000 record i need to search in. Every week there are 500 000 records added.
so, this is what I need:
location_id value date time name lat lng
3 234 2011-11-18 19:50:00 Amerongen beneden 5.40453 51.97486
4 594 2011-11-18 19:50:00 Amerongen boven 5.41194 51.97507
I do this with this query:
SELECT location_id, value, date, time, locations.name, locations.lat, locations.lng FROM
(
SELECT location_id, value, date, time from `measurements`
LEFT JOIN units ON (units.id = measurements.unit_id)
WHERE units.name='Waterhoogte'
ORDER BY measurements.date DESC, measurements.time DESC
) as last_record
LEFT JOIN locations on (locations.id = location_id)
GROUP BY location_id
which takes 30 seconds. How can I improve this? This is my structure:
CREATE TABLE IF NOT EXISTS `locations` (
`id` int(11) NOT NULL auto_increment,
`name` varchar(255) NOT NULL,
`code` varchar(255) NOT NULL,
`lat` varchar(10) NOT NULL,
`lng` varchar(10) NOT NULL,
`owner_id` int(11) NOT NULL,
PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=244 ;
-- --------------------------------------------------------
--
-- Table structure for table `measurements`
--
CREATE TABLE IF NOT EXISTS `measurements` (
`id` int(11) NOT NULL auto_increment,
`date` date NOT NULL,
`time` time NOT NULL,
`value` varchar(255) NOT NULL,
`location_id` int(11) NOT NULL,
`unit_id` int(11) NOT NULL,
PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=676801 ;
-- --------------------------------------------------------
--
-- Table structure for table `owner`
--
CREATE TABLE IF NOT EXISTS `owner` (
`id` int(11) NOT NULL auto_increment,
`name` varchar(255) NOT NULL,
PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=3 ;
-- --------------------------------------------------------
--
-- Table structure for table `units`
--
CREATE TABLE IF NOT EXISTS `units` (
`id` int(11) NOT NULL auto_increment,
`name` varchar(255) NOT NULL,
`description` text NOT NULL,
`unit_short` varchar(255) NOT NULL,
`owner_id` int(11) NOT NULL,
PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=44 ;
What is the limit what phpmyadmin can handle?

Answer 1

Create an index on units.name specifically is a good start.
You should also really rethink the amount of data you are pulling back.
Is someone really going to sift through that many records. Change your query to limit the number of records and think of a UI interface that involves a paging mechanism.

Answer 2

you need to put an index or unique index on units.name.

Answer 3

Add the following indexes:
A composite index (a covering index) on unit.name and unit.id.
A composite index of measurements.date and measurements.time.
An index on location.id

Answer 4

You should try creating an index on units.name as a first step. But understand that there is a tradeoff with an index - read operations will be faster, but it can slow down write operations. If you're concerned about that, or if you're affected by slow writes, then you may want to try creating the index on a smaller number of characters in units.name.
For instance, to declare an index on the first 12 characters of units.name, you'd declare the following:
CREATE INDEX first_twelve ON units (name(12));
Again, this may not be necessary if you don't notice any ill effects from just throwing an index on, but it's something to keep in mind.

Answer 5

SELECT measurements.location_id, measurements.value, measurements.date, measurements.time, locations.name, locations.lat, locations.lng
FROM measurements
LEFT JOIN units ON units.id = measurements.unit_id
LEFT JOIN locations ON locations.id = measurements.location_id
WHERE units.id = 4
GROUP BY measurements.location_id
ORDER BY measurements.date DESC, measurements.time DESC