select from multiple tables but ordering by a datetime field - mysql

Question

I have 3 tables that are unrelated (related that each contains data for a different social network). Each has a datetime field dated- I'm already grouping by hour as you can see below (this one below for linked_in)
SELECT count(*), date_format(dated, '%Y:%m:%d %H') as hour
FROM upd8r_linked_in_accts
WHERE CAST(dated AS DATE) = '".$start_date."'
GROUP BY hour
I would like to know how to do a total across all 3 networks- the tables for the three are
CREATE TABLE IF NOT EXISTS `upd8r_facebook_accts` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`owner_id` varchar(50) NOT NULL,
`user_id` int(11) NOT NULL,
`fb_id` bigint(30) NOT NULL,
`dated` datetime NOT NULL,
PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=80 ;
CREATE TABLE IF NOT EXISTS `upd8r_linked_in_accts` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`owner_id` varchar(50) NOT NULL,
`user_id` int(11) NOT NULL,
`linked_in` varchar(200) NOT NULL,
`oauth_secret` varchar(100) NOT NULL,
`first_count` int(11) NOT NULL,
`second_count` int(11) NOT NULL,
`dated` datetime NOT NULL,
PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=200 ;
CREATE TABLE IF NOT EXISTS `upd8r_twitter_accts` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`owner_id` varchar(50) NOT NULL,
`user_id` int(11) NOT NULL,
`twitter` varchar(200) NOT NULL,
`twitter_secret` varchar(100) NOT NULL,
`dated` datetime NOT NULL,
PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=9 ;
something like this ?
(SELECT count(*), date_format(dated, '%Y:%m:%d %H') as hour
FROM upd8r_linked_in_accts
WHERE CAST(dated AS DATE) = '".$start_date."')
UNION ALL
(SELECT count(*), date_format(dated, '%Y:%m:%d %H') as hour
FROM upd8r_facebook_accts
WHERE CAST(dated AS DATE) = '".$start_date."')
UNION ALL
(SELECT count(*), date_format(dated, '%Y:%m:%d %H') as hour
FROM upd8r_twitter_accts
WHERE CAST(dated AS DATE) = '".$start_date."')
UNION ALL
GROUP BY hour
update
the data in the actual database is not related by any pk or fk but the data is related that each table represents a user registering for a social network through the application and therefore i need to show the amount of users registered per hour across all three tables
update 2
the output of the query should show %Y:%m:%d %H and the amount of users registered (records created) that hour across the three tables.. each hour returning a new row (ordered by the time)

Answer 1

You are close to the solution
select t1.hourx, sum(t1.column1)
from (
(
SELECT count(*) as column1, date_format(dated, '%Y:%m:%d %H') as hourx
FROM upd8r_linked_in_accts
WHERE CAST(dated AS DATE) = '".$start_date."'
GROUP BY hourx
)
UNION ALL
(
SELECT count(*) as column1, date_format(dated, '%Y:%m:%d %H') as hourx
FROM upd8r_facebook_accts
WHERE CAST(dated AS DATE) = '".$start_date."'
GROUP BY hourx
)
UNION ALL
(
SELECT count(*) as column1, date_format(dated, '%Y:%m:%d %H') as hourx
FROM upd8r_twitter_accts
WHERE CAST(dated AS DATE) = '".$start_date."'
GROUP BY hourx
)
) t1
GROUP BY t1.hourx
I use 'hourx' to avoid reserve words, maybe not necessary.
I hope this works.
[ADD] This solution is called 'inline view'. You can google for that. It is supported by most of databases (mysql, oracle mssql and etc)