MySQL inner join issue selecting max row - mysql

I have some SQL below the problem I have is both leads (cosmic_leads) show with the same lead stage. I'm finding it hard to explain so I will post a SQL dump at the bottom so you can see table structures etc.
Here is the SQL:-
SELECT
cl.*,
clt.stage_name AS stagename,
cls.date AS moddate,
cls.time AS modtime,
cls.comments
FROM
cosmic_leads cl
INNER JOIN
cosmic_leads_stages cls
ON
cls.lead_id = cl.id
INNER JOIN (
SELECT MAX(id) as id
FROM cosmic_leads_stages
GROUP BY site, lead_id
) clsid
ON
cls.id = clsid.id
INNER JOIN
cosmic_leads_types clt
ON
clt.type = cl.type AND clt.site = cl.site AND clt.stage = cls.stage
WHERE
cls.date >= 20140701 AND cls.date <= 20140723
GROUP BY
cl.id, cl.site
ORDER BY
cls.date DESC LIMIT 0, 10
Using this SQL I get the following:-
[
{
'modtime' => '145717',
'sale' => '1',
'name' => undef,
'moddate' => '20140723',
'comments' => undef,
'username' => 'aap',
'site' => '1',
'handler' => undef,
'stagename' => 'Closed - Success',
'id' => 'lead1',
'type' => 'Website'
},
{
'modtime' => '145717',
'sale' => '0',
'name' => undef,
'moddate' => '20140723',
'comments' => undef,
'username' => 'aap',
'site' => '2',
'handler' => undef,
'stagename' => 'Closed - Success',
'id' => 'lead1',
'type' => 'Website'
}
];
The stage name for lead1 on site 2 should read "New", can anyone see what the issue is?
Thanks
(SQL dump below)
--
-- Table structure for table cosmic_leads
CREATE TABLE IF NOT EXISTS `cosmic_leads` (
`id` varchar(128) NOT NULL,
`type` varchar(96) NOT NULL,
`site` int(3) NOT NULL,
`username` varchar(96) NOT NULL,
`sale` varchar(48) NOT NULL DEFAULT '0',
`name` varchar(96) DEFAULT NULL,
`handler` varchar(96) DEFAULT NULL,
PRIMARY KEY (`id`,`site`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1;
--
-- Dumping data for table cosmic_leads
INSERT INTO `cosmic_leads` (`id`, `type`, `site`, `username`, `sale`, `name`, `handler`) VALUES
('lead1', 'Website', 1, 'aap', '1', NULL, NULL),
('lead1', 'Website', 2, 'aap', '0', NULL, NULL);
--
-- Table structure for table cosmic_leads_stages
CREATE TABLE IF NOT EXISTS `cosmic_leads_stages` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`lead_id` varchar(128) NOT NULL,
`site` int(3) NOT NULL,
`stage` int(3) NOT NULL,
`date` int(8) NOT NULL,
`time` int(6) NOT NULL,
`comments` varchar(128) DEFAULT NULL,
`comments_internal` varchar(128) DEFAULT NULL,
`extra` varchar(128) DEFAULT NULL,
PRIMARY KEY (`id`),
UNIQUE KEY `lead_id_site_stage` (`lead_id`,`site`,`stage`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1 AUTO_INCREMENT=11 ;
--
-- Dumping data for table cosmic_leads_stages
INSERT INTO `cosmic_leads_stages` (`id`, `lead_id`, `site`, `stage`, `date`, `time`, `comments`, `comments_internal`, `extra`) VALUES
(8, 'lead1', 1, 1, 20140723, 145701, NULL, NULL, NULL),
(9, 'lead1', 1, 9, 20140723, 145717, NULL, NULL, NULL),
(10, 'lead1', 2, 1, 20140723, 145724, NULL, NULL, NULL);
--
-- Table structure for table cosmic_leads_types
CREATE TABLE IF NOT EXISTS `cosmic_leads_types` (
`id` int(12) NOT NULL AUTO_INCREMENT,
`type` varchar(128) NOT NULL,
`site` int(3) NOT NULL,
`stage` int(3) NOT NULL,
`stage_name` varchar(128) NOT NULL,
`status` int(2) NOT NULL,
`amount` decimal(10,2) NOT NULL DEFAULT '0.00',
PRIMARY KEY (`id`),
UNIQUE KEY `type` (`type`,`site`,`stage`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1 AUTO_INCREMENT=13 ;
--
-- Dumping data for table cosmic_leads_types
INSERT INTO `cosmic_leads_types` (`id`, `type`, `site`, `stage`, `stage_name`, `status`, `amount`) VALUES
(4, 'Website', 1, 1, 'New', 1, '0.00'),
(5, 'Website', 1, 9, 'Closed - Success', 4, '0.00'),
(6, 'Website', 1, 10, 'Closed - Failed', 5, '0.00'),
(10, 'Website', 2, 1, 'New', 1, '0.00'),
(11, 'Website', 2, 9, 'Closed - Success', 4, '0.00'),
(12, 'Website', 2, 10, 'Closed - Failed', 5, '0.00');
--
-- Constraints for dumped tables
--
-- Constraints for table cosmic_leads_stages
ALTER TABLE `cosmic_leads_stages`
ADD CONSTRAINT `cosmic_leads_stages_ibfk_1` FOREIGN KEY (`lead_id`, `site`) REFERENCES `cosmic_leads` (`id`, `site`) ON DELETE CASCADE;

You have 2 cosmic_leads records, each of which have 3 records on cosmic_leads_stages.
Those 6 are joined to the max id of each site / lead_id from cosmic_leads_stages, bringing it down to 4 records. 2 for each record on cosmic_leads:-
ID TYPE SITE USERNAME SALE NAME HANDLER MODDATE MODTIME COMMENTS STAGE
lead1 Website 2 aap 0 (null) (null) 20140723 145724 (null) 1
lead1 Website 1 aap 1 (null) (null) 20140723 145724 (null) 1
lead1 Website 2 aap 0 (null) (null) 20140723 145717 (null) 9
lead1 Website 1 aap 1 (null) (null) 20140723 145717 (null) 9
You then join those 4 records against cosmic_leads_types based on type, site and stage and all 4 have a match, resulting in 4 records.
You then use GROUP BY cl.id, cl.site to reduce this to 1 records for each id / site. For each id / site it has found a match on Closed - Success and on New . The GROUP BY forces it to randomly pick one of these in each case (which one it choses is not defined).
As you current query goes it is producing what is to be expected.
However I think to get the results I think you want you need to include site in the join of cosmic_leads and cosmic_leads_stages
SELECT
cl.*,
clt.stage_name AS stagename,
cls.date AS moddate,
cls.time AS modtime,
cls.comments
FROM cosmic_leads cl
INNER JOIN cosmic_leads_stages cls
ON cls.lead_id = cl.id
AND cls.site = cl.site
INNER JOIN
(
SELECT MAX(id) as id, site, lead_id
FROM cosmic_leads_stages
GROUP BY site, lead_id
) clsid
ON cls.id = clsid.id
AND cls.site = clsid.site
AND cls.lead_id = clsid.lead_id
INNER JOIN cosmic_leads_types clt
ON clt.type = cl.type AND clt.site = cl.site AND clt.stage = cls.stage
WHERE cls.date >= 20140701 AND cls.date <= 20140723
ORDER BY cls.date DESC LIMIT 0, 10

Related

Else is not running in mysql Query

I have below query. In this I have yes and no case.
yes is accessing but else part is not working . Please have a look on this.
SELECT SalesChannel.name , count(Transaction.category_id) as count, (case when (Transaction.no_of_units > 0 and Transaction.mop > 0) THEN 'yes' ELSE 'No' END) AS Is_Present from outlets Outlet inner join transactions Transaction on Outlet.id = Transaction.outlet_id inner join sale_channels SalesChannel on SalesChannel.id = Outlet.sale_channel_id group by SalesChannel.name
the output should be as below
KU Electrical
Yes 6 2
No 1 2
6 is counter of KU and Yes refers the presence,similarly No is non presence of KU
select SalesChannel.name ,
Transaction.category_id,
count(Transaction.category_id) as count,
from outlets Outlet inner join transactions Transaction on Outlet.id = Transaction.outlet_id inner join sale_channels SalesChannel on SalesChannel.id = Outlet.sale_channel_id group by SalesChannel.name
below are three tables which i used
1. transactions
CREATE TABLE IF NOT EXISTS `transactions` (
`id` int(11) NOT NULL,
`zone_id` int(11) NOT NULL,
`state_id` int(11) NOT NULL,
`city_id` int(11) NOT NULL,
`category_id` int(11) NOT NULL,
`sub_category_id` int(11) NOT NULL,
`brand_id` int(11) NOT NULL,
`model_id` int(11) NOT NULL,
`outlet_id` int(11) NOT NULL,
`no_of_units` int(11) NOT NULL,
`mop` decimal(10,2) NOT NULL
) ENGINE=InnoDB AUTO_INCREMENT=5 DEFAULT CHARSET=latin1;
--
-- Dumping data for table `transactions`
--
INSERT INTO `transactions` (`id`, `zone_id`, `state_id`, `city_id`, `category_id`, `sub_category_id`, `brand_id`, `model_id`, `outlet_id`, `no_of_units`, `mop`) VALUES
(1, 2, 2, 2, 2, 1, 1, 1, 1, 3, '6.00'),
(2, 2, 2, 2, 2, 1, 1, 1, 1, 3, '6.00'),
(3, 1, 1, 1, 1, 1, 1, 1, 1, 4, '2.00'),
(4, 2, 2, 2, 1, 1, 1, 1, 2, 4, '2.00');
2.outlets
CREATE TABLE IF NOT EXISTS `outlets` (
`id` int(11) NOT NULL,
`outlet_code` varchar(255) NOT NULL,
`name` varchar(255) NOT NULL,
`zone_id` int(11) NOT NULL,
`state_id` int(11) NOT NULL,
`city_id` int(11) NOT NULL,
`sale_channel_id` int(11) NOT NULL,
`is_active` tinyint(1) NOT NULL,
`created` datetime NOT NULL,
`modified` datetime NOT NULL
) ENGINE=InnoDB AUTO_INCREMENT=3 DEFAULT CHARSET=latin1;
--
-- Dumping data for table `outlets`
--
INSERT INTO `outlets` (`id`, `outlet_code`, `name`, `zone_id`, `state_id`, `city_id`, `sale_channel_id`, `is_active`, `created`, `modified`) VALUES
(1, '1508', 'Ashok electricals', 2, 2, 2, 1, 1, '2016-10-03 00:00:00', '2016-10-03 00:00:00'),
(2, '1233', 'vinayak electricals', 1, 1, 1, 2, 1, '2016-10-04 00:00:00', '2016-10-04 00:00:00');
3. sale_chennals
CREATE TABLE IF NOT EXISTS `sale_channels` (
`id` int(11) NOT NULL,
`name` varchar(255) NOT NULL,
`is_active` tinyint(1) NOT NULL,
`created` datetime NOT NULL,
`modified` datetime NOT NULL
) ENGINE=InnoDB AUTO_INCREMENT=3 DEFAULT CHARSET=latin1;
--
-- Dumping data for table `sale_channels`
--
INSERT INTO `sale_channels` (`id`, `name`, `is_active`, `created`, `modified`) VALUES
(1, 'KU', 1, '2016-10-03 00:00:00', '2016-10-03 00:00:00'),
(2, 'Electricals', 1, '2016-10-04 00:00:00', '2016-10-04 00:00:00');
There is no data in tables that match for the else condition. Your condition is that "Transaction.no_of_units >0 AND Transaction.mop >0" that is not match in table value of both fields are greater than 0.
Otherwise, else condition works fine.
You are aggregating your data so as to get one row per SalesChannel.name. There may be some transaction records that result in 'Yes' and others in 'No' for a SalesChannel.name, so what then is Is_Present supposed to be?
Another issue with your query is that the sale channels are in a table. There are currently two of them, but there could be three or four or thousands sometime. A SQL query doesn't produce a result with a variable number of columns. The columns must be known beforehand. So a possible result could look like this:
Name Yes No
KU 6 1
Electrical 2 2
because you know you want it to be Yes or No only, no matter how many channels.
The query:
select
sc.name,
count(case when t.no_of_units > 0 and t.mop > 0 then 1 end) as yes,
count(case when t.no_of_units <= 0 or t.mop <= 0 then 1 end) as no
from sale_channels sc
join outlet o on o.sale_channel_id = sc.id
join transactions t on t.outlet_id = o.id;

showing counter in mysql query

I have mysql query given below. I which counter has been used. if i enter category Id 1 for 3 times then counter is coming 3 which is correct but with this i want if i do not enter then different coloumn should come with NO.
output should be
KU Electrical
Yes 6 2
No 1 2
In this KU and Electrical are my sale channel name. Yes means counter of enteries of KU and No means which have not entered. Please help out in this. i am struggling
select
SalesChannel.name,
Transaction.category_id,
count(Transaction.category_id) as "count"
from outlets Outlet
inner join transactions Transaction on Outlet.id = Transaction.outlet_id
inner join sale_channels SalesChannel on SalesChannel.id = Outlet.sale_channel_id
group by Transaction.category_id;
below are three tables which I used
1) transactions
CREATE TABLE IF NOT EXISTS `transactions` (
`id` int(11) NOT NULL,
`zone_id` int(11) NOT NULL,
`state_id` int(11) NOT NULL,
`city_id` int(11) NOT NULL,
`category_id` int(11) NOT NULL,
`sub_category_id` int(11) NOT NULL,
`brand_id` int(11) NOT NULL,
`model_id` int(11) NOT NULL,
`outlet_id` int(11) NOT NULL,
`no_of_units` int(11) NOT NULL,
`mop` decimal(10,2) NOT NULL
) ENGINE=InnoDB AUTO_INCREMENT=5 DEFAULT CHARSET=latin1;
--
-- Dumping data for table `transactions`
--
INSERT INTO `transactions` (`id`, `zone_id`, `state_id`, `city_id`, `category_id`, `sub_category_id`, `brand_id`, `model_id`, `outlet_id`, `no_of_units`, `mop`) VALUES
(1, 2, 2, 2, 2, 1, 1, 1, 1, 3, '6.00'),
(2, 2, 2, 2, 2, 1, 1, 1, 1, 3, '6.00'),
(3, 1, 1, 1, 1, 1, 1, 1, 1, 4, '2.00'),
(4, 2, 2, 2, 1, 1, 1, 1, 2, 4, '2.00');
2) outlets
CREATE TABLE IF NOT EXISTS `outlets` (
`id` int(11) NOT NULL,
`outlet_code` varchar(255) NOT NULL,
`name` varchar(255) NOT NULL,
`zone_id` int(11) NOT NULL,
`state_id` int(11) NOT NULL,
`city_id` int(11) NOT NULL,
`sale_channel_id` int(11) NOT NULL,
`is_active` tinyint(1) NOT NULL,
`created` datetime NOT NULL,
`modified` datetime NOT NULL
) ENGINE=InnoDB AUTO_INCREMENT=3 DEFAULT CHARSET=latin1;
--
-- Dumping data for table `outlets`
--
INSERT INTO `outlets` (`id`, `outlet_code`, `name`, `zone_id`, `state_id`, `city_id`, `sale_channel_id`, `is_active`, `created`, `modified`) VALUES
(1, '1508', 'Ashok electricals', 2, 2, 2, 1, 1, '2016-10-03 00:00:00', '2016-10-03 00:00:00'),
(2, '1233', 'vinayak electricals', 1, 1, 1, 2, 1, '2016-10-04 00:00:00', '2016-10-04 00:00:00');
3) sale_chennals
CREATE TABLE IF NOT EXISTS `sale_channels` (
`id` int(11) NOT NULL,
`name` varchar(255) NOT NULL,
`is_active` tinyint(1) NOT NULL,
`created` datetime NOT NULL,
`modified` datetime NOT NULL
) ENGINE=InnoDB AUTO_INCREMENT=3 DEFAULT CHARSET=latin1;
--
-- Dumping data for table `sale_channels`
--
INSERT INTO `sale_channels` (`id`, `name`, `is_active`, `created`, `modified`) VALUES
(1, 'KU', 1, '2016-10-03 00:00:00', '2016-10-03 00:00:00'),
(2, 'Electricals', 1, '2016-10-04 00:00:00', '2016-10-04 00:00:00');
SQL fiddle: http://sqlfiddle.com/#!9/3f497/1
You are grouping by category. That means you get one result row per category. In each of these rows you show the count and a sale channel name. This sale channel name is just one of the names found in the records for the category arbitrarily chosen.
I suppose you want to count per category and sale channel. Hence your group by clause should be group by SalesChannel.name, Transaction.category_id:
select
SalesChannel.name,
Transaction.category_id,
count(Transaction.category_id) as "count"
from outlets Outlet
inner join transactions Transaction on Outlet.id = Transaction.outlet_id
inner join sale_channels SalesChannel on SalesChannel.id = Outlet.sale_channel_id
group by SalesChannel.name, Transaction.category_id;
SQL fiddle: http://sqlfiddle.com/#!9/3f497/2
This result, however, doesn't show an entry for Electricals / category 2, because there is no transaction for this combination in the table. If you want to show a zero count for this, you'd have to create the complete result set (i.e. all combinations of channel and category, whether they have a transaction or not) first. Then you'd outer join the transactions:
select
sc.name,
c.id as category_id,
count(t.id) as "count"
from sale_channels sc
cross join categories c
left join outlets o on o.sale_channel_id = sc.id
left join transactions t on t.outlet_id = o.id and t.category_id = c.id
group by sc.name, c.id;
SQL fiddle: http://sqlfiddle.com/#!9/60e998/5

MySQL - Correct approach event counting

I want to list users which have a particular event count but I'm confused on which approach to take.
This is the database table:
CREATE TABLE `event` (
`event_id` int(11) unsigned NOT NULL AUTO_INCREMENT,
`visitor_id` int(11) DEFAULT NULL,
`key` varchar(200) DEFAULT NULL,
`value` text,
`label` varchar(200) DEFAULT '',
`datetime` datetime DEFAULT NULL,
PRIMARY KEY (`event_id`)
) ENGINE=InnoDB AUTO_INCREMENT=6 DEFAULT CHARSET=utf8;
INSERT INTO `event` (`event_id`, `visitor_id`, `key`, `value`, `label`, `datetime`)
VALUES
(1, 1, 'LOGIN', NULL, '', NULL),
(2, 2, 'LOGIN', NULL, '', NULL),
(3, 1, 'VIEW_PAGE', 'HOTEL', '', NULL),
(4, 2, 'VIEW_PAGE', 'HOTEL', '', NULL),
(5, 1, 'PURCHASE_HOTEL', NULL, '', NULL);
CREATE TABLE `visitor` (
`visitor_id` int(11) unsigned NOT NULL AUTO_INCREMENT,
`datetime` datetime DEFAULT NULL,
PRIMARY KEY (`visitor_id`)
) ENGINE=InnoDB AUTO_INCREMENT=6 DEFAULT CHARSET=utf8;
INSERT INTO `visitor` (`visitor_id`, `datetime`)
VALUES
(1, NULL),
(2, NULL);
and this is my approach:
SELECT DISTINCT
t1.`visitor_id`
FROM
`visitor` t1
JOIN `event` t2 on t1.visitor_id = t2.visitor_id AND t2.`key` = 'LOGIN'
JOIN `event` t3 on t1.visitor_id = t3.visitor_id AND t3.`key` = 'VIEW_PAGE' AND t3.`value` = 'HOTEL'
WHERE ( SELECT COUNT(*) FROM `event` WHERE `event`.`key` = 'PURCHASE_HOTEL' ) > 0
this should only list visitor 1 but it does actually list visitor 2 too which does not have the PURCHASE_HOTEL event.
As you can imagine, there will be more "rules" like all the JOIN events for each particular case. Can we correct and improve this somehow?
BONUS:
What is the name of this approach?
I think this is a "set-within-sets" query. I like using aggregation with a having clause for this type of query. The following checks the three conditions you are looking for:
select visitor_id
from event e
group by visitor_id
having sum(e.key = 'LOGIN') > 0 and
sum(e.key = 'VIEW_PAGE' and e.value = 'HOTEL') > 0 and
sum(e.key = 'PURCHASE_HOTEL') > 0;
The first condition in the having clause counts the number of LOGIN records and is true when at least one is found. (If you want exactly one, change > 0 to = 0.)
The second condition checks the viewing of the hotel page.
The third counts the number of hotel purchases.

right join query returning null value records

CREATE TABLE IF NOT EXISTS `tweets_comment_tbl` (
`tweet_comment_id` int(12) NOT NULL AUTO_INCREMENT,
`tweet_id` int(12) NOT NULL,
`tweets_comment` text NOT NULL,
`created` int(12) NOT NULL,
`changed` int(12) NOT NULL,
`uid` int(12) NOT NULL,
`userip` varchar(20) NOT NULL,
`referer` text NOT NULL,
`status` int(2) NOT NULL DEFAULT '1',
PRIMARY KEY (`tweet_comment_id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=6 ;
--
-- Dumping data for table `tweets_comment_tbl`
--
INSERT INTO `tweets_comment_tbl` (`tweet_comment_id`, `tweet_id`, `tweets_comment`, `created`, `changed`, `uid`, `userip`, `referer`, `status`) VALUES
(1, 1, 'COMMENT USER1', 1319395671, 1319395671, 3, '127.0.0.1', 'http://localhost/drupal_tutorial/', 1),
(2, 2, 'comment admin user', 1319395724, 1319395724, 1, '127.0.0.1', 'http://localhost/drupal_tutorial/node', 1),
(3, 2, 'USER COMMENTING HIS COMMENT', 1319395838, 1319395838, 3, '127.0.0.1', 'http://localhost/drupal_tutorial/', 1),
(4, 2, 'ADMIN COMMENTING FOR HIS COMMENT', 1319395865, 1319395865, 1, '127.0.0.1', 'http://localhost/drupal_tutorial/node', 1),
(5, 2, 'dddCOMMENT USER1: ADMIN DOING COMMENT FOR STATUS UPDATE1', 1319395905, 1319395905, 1, '127.0.0.1', 'http://localhost/drupal_tutorial/node', 1);
my second table
CREATE TABLE IF NOT EXISTS `tweets_tbl` (
`tweet_id` int(11) NOT NULL AUTO_INCREMENT,
`tweets` text NOT NULL,
`created` int(12) NOT NULL,
`changed` int(12) NOT NULL,
`uid` int(12) NOT NULL,
`userip` varchar(20) NOT NULL,
`referer` text NOT NULL,
`status` int(2) NOT NULL DEFAULT '1',
PRIMARY KEY (`tweet_id`),
KEY `tweet_id` (`tweet_id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=3 ;
--
-- Dumping data for table `tweets_tbl`
--
INSERT INTO `tweets_tbl` (`tweet_id`, `tweets`, `created`, `changed`, `uid`, `userip`, `referer`, `status`) VALUES
(1, 'STATUS UPDATE 1', 1319395633, 1319395633, 1, '127.0.0.1', 'http://localhost/drupal_tutorial/node', 1),
(2, 'Status update user1', 1319395696, 1319395696, 3, '127.0.0.1', 'http://localhost/drupal_tutorial/node', 1);
Trying join query
SELECT ut.picture as picture, tct.tweet_id as tweet_id, tct.tweets_comment as tweets_comment, tct.changed
FROM tweets_comment_tbl tct
RIGHT JOIN users as ut ON tct.uid=ut.uid AND tct.tweet_id=2 AND tct.status = 1
order by tct.created desc
return records for above query
picture tweet_id tweets_comment changed
1 COMMENT USER1 1319395671
NULL NULL NULL
picture-1.png NULL NULL NULL
actually what i expected is
picture tweet_id tweets_comment changed
1 COMMENT USER1 1319395671
Why query has been returning NULL records, i am not sure where i made mistake in join query.
The query returns all rows from users and joins tweets_comment_tbl based on condition in ON. If no records in tweets_comment_tbl matches userid, the record from users still included into recordset. You probably need INNER JOIN if you want to see just users with comments.
Side note :
I think it's almost always possible to avoid RIGHT JOIN; queries with LEFT JOIN are much easier to read and understand.

Dynamic Foreign Keys - How To Implement?

I have 4 tables (appointed, class, elected, status) that I want to cross reference into a single table's (members) column. The values of the of 4 tables are time sensitive based off a history table (members_history). The desired result is that the query should output all members and the current appointed position or current elected position, class, and status within the members row and include additional information obtained from the foreign rows.
So instead of just returning:
id, username, password, salt, name_first, name_last, date_join & date_leave;
The query would return
id, username, password, salt, name_prefix, name_first, name_last, hours_extra, date_join, date_leave, appointed, class, elected & status;
Wherever an added column does not have a current value in history it's result should be NULL.
Now I think I can do this with sub-querys, but have been so far banging my head against the keyboard. I'll take another swing at it later, but until then, anyone else willing to give it a shot, or attempt to point me in the right direction?
The structure of my SQL (no pun intended) tables is as follows:
CREATE TABLE IF NOT EXISTS `members` (
`id` mediumint(3) unsigned NOT NULL auto_increment COMMENT 'Members Unique Id',
`username` varchar(32) collate utf8_bin NOT NULL COMMENT 'Mebers Username',
`password` varchar(64) collate utf8_bin NOT NULL COMMENT 'Members Password Hash',
`salt` varchar(32) collate utf8_bin NOT NULL COMMENT 'Members Password Salt',
`name_first` varchar(32) collate utf8_bin NOT NULL COMMENT 'Members First Name',
`name_last` varchar(32) collate utf8_bin NOT NULL COMMENT 'Members Last Name',
`date_join` date NOT NULL COMMENT 'Members Join Date',
`date_leave` date default NULL COMMENT 'Members Resgination Date (If Applicable)',
PRIMARY KEY (`id`),
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_bin COMMENT='Members id in this table = mid in other tables';
CREATE TABLE IF NOT EXISTS `members:apointed` (
`id` tinyint(3) unsigned NOT NULL auto_increment COMMENT 'Unique value',
`name_prefix` varchar(8) collate utf8_bin NOT NULL COMMENT 'Prefix Added to Members Name',
`hours_extra` decimal(4,2) NOT NULL COMMENT 'Hours Given as Bonus for Holding this Position.',
`position` varchar(40) collate utf8_bin NOT NULL COMMENT 'Name of the Posisiton',
PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_bin COMMENT='Undefined within the SOP or By-Laws.';
CREATE TABLE IF NOT EXISTS `members:class` (
`id` tinyint(3) unsigned NOT NULL auto_increment COMMENT 'Unique Id',
`class` varchar(8) collate utf8_bin NOT NULL COMMENT 'Unique Value',
PRIMARY KEY (`id`),
UNIQUE KEY `value` (`class`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_bin COMMENT='Article I, Section 1 Subsection B: Classes of Membership';
CREATE TABLE IF NOT EXISTS `members:elected` (
`id` tinyint(3) unsigned NOT NULL auto_increment COMMENT 'Unique value',
`name_prefix` varchar(8) collate utf8_bin NOT NULL COMMENT 'Prefix Added to Members Name',
`hours_extra` decimal(4,2) NOT NULL COMMENT 'Hours Given as Bonus for Holding this Position.',
`position` varchar(40) collate utf8_bin NOT NULL COMMENT 'Name of the Posisiton',
PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_bin COMMENT='Article II';
CREATE TABLE IF NOT EXISTS `members:status` (
`id` tinyint(3) unsigned NOT NULL auto_increment COMMENT 'Bit''s Place',
`status` varchar(16) collate utf8_bin NOT NULL COMMENT 'Categorie''s Name',
PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_bin COMMENT='Article I, Section 1, Subsection A: Categories of Membership';
CREATE TABLE IF NOT EXISTS `members_history` (
`id` int(10) unsigned NOT NULL auto_increment COMMENT 'Unique Id',
`mid` tinyint(3) unsigned NOT NULL COMMENT 'Members Unique Id.',
`table` enum('class','elected','appointed','status') NOT NULL COMMENT 'Name of Table that was Edited.',
`value` tinyint(3) unsigned NOT NULL COMMENT 'Value',
`start` date NOT NULL COMMENT 'Value''s Effect Date',
`end` date default NULL COMMENT 'Value''s Expiration Date',
PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1 COMMENT='Member History';
members_history.mid is a FK for id in the members table, not every member will have history on them (but eventually they all will, as every member will have to have a class and status). members_history.value is a FK for members:{members_history.table}.id;
INSERT INTO `members`
(`id`, `username`, `password`, `salt`, `name_first`, `name_last`, `date_join`, `date_join`) VALUES
( 1, 'Dygear',MD5('pass'), 's417', 'Mark', 'Tomlin', DATE(), NULL),
( 2, 'uberusr',MD5('p455'), '235f', 'Howard', 'Singer', DATE(), NULL),
( 3,'kingchief',MD5('leet'), '32fs','Christopher', 'Buckham', DATE(), NULL);
INSERT INTO `members:apointed`
(`id`, `name_prefix`, `hours_extra`, `posisiton`) VALUES
( 1, '', 0.00, 'Crew Chief'),
( 2, '', 20.00, 'Engineer'),
( 3, 'Lt.', 40.00, 'Lieutenant'),
( 4, 'Capt.', 60.00, 'Captin'),
( 5, 'Chief.', 80.00, '3rd Assistant Chief of Operation');
INSERT INTO `members:class`
(`id`, `class`) VALUES
( 1, 'Class I'),
( 2, 'Class II');
INSERT INTO `members:elected`
(`id`, `name_prefix`, `hours_extra`, `posisiton`) VALUES
( 1, '', 40.00, 'Trustee'),
( 2, '', 40.00, 'Chairman of the Board'),
( 3, 'Prez.', 40.00, 'President'),
( 4, 'VPrez.', 40.00, 'Vice-President'),
( 5, '', 40.00, 'Recording Secretary'),
( 6, '', 40.00, 'Service Secretary'),
( 7, '', 40.00, 'Corresponding Secretary'),
( 8, '', 40.00, 'Financial Secretary Treasuer'),
( 9, '', 40.00, 'Assistant Financial Secretary Treasuer'),
( 10, 'Chief.', 80.00, 'Chief of Operations'),
( 11, 'Chief.', 80.00, 'First Deputy Chief of Operations'),
( 12, 'Chief.', 80.00, 'Second Deputy Chief of Operation');
INSERT INTO `members:status`
(`id`, `status`) VALUES
( 1, 'Active'),
( 2, 'Inactive'),
( 3, 'Student'),
( 4, 'Probationary'),
( 5, 'Lifetime'),
( 6, 'Cadet'),
( 7, 'Honorary'),
( 8, 'Medical'),
( 9, 'Military'),
( 10, 'Resigned'),
( 11, 'Disvowed');
INSERT INTO `members_history`
(`id`, `mid`, `table`, `value`, `start`, `end`) VALUES
(NULL, 1, 'apointed', 3, DATE(), NULL),
(NULL, 1, 'class', 1, DATE(), NULL),
(NULL, 1, 'status', 1, DATE(), NULL),
(NULL, 2, 'elected', 4, DATE(), NULL),
(NULL, 2, 'class', 1, DATE(), NULL),
(NULL, 2, 'status', 1, DATE(), NULL),
(NULL, 3, 'apointed', 10, DATE(), '2010-05-01'),
(NULL, 3, 'class', 1, DATE(), NULL),
(NULL, 3, 'status', 1, DATE(), NULL);
You're using a design called polymorphic associations and it's frequently done wrong. The way to make it work is to create another table, say members:abstract:
CREATE TABLE IF NOT EXISTS `members:abstract` (
`id` INT UNSIGNED AUTO_INCREMENT PRIMARY KEY,
`type` enum('class','elected','appointed','status') NOT NULL,
UNIQUE KEY (`id`, `type`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_bin;
This table serves as the parent table for all of your members attributes tables. Each of these tables changes its primary key to not generate id values automatically, but instead reference the primary key of members:abstract. I'll show just members:appointed but the others would be similar.
CREATE TABLE IF NOT EXISTS `members:appointed` (
`id` INT UNSIGNED NOT NULL PRIMARY KEY, -- not auto_increment
`name_prefix` varchar(8) collate utf8_bin NOT NULL COMMENT 'Prefix Added to Members Name',
`hours_extra` decimal(4,2) NOT NULL COMMENT 'Hours Given as Bonus for Holding this Position.',
`position` varchar(40) collate utf8_bin NOT NULL COMMENT 'Name of the Posisiton',
FOREIGN KEY (`id`) REFERENCES `members:abstract` (`id`) ON DELETE CASCADE
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_bin COMMENT='Undefined within the SOP or By-Laws.';
You can make this table gain auto-generated values automatically with a trigger:
DELIMITER //
DROP TRIGGER IF EXISTS ins_appointed//
CREATE TRIGGER ins_appointed BEFORE INSERT ON `members:appointed`
FOR EACH ROW BEGIN
INSERT INTO `members:abstract` (`type`) VALUES ('appointed');
SET NEW.id = LAST_INSERT_ID();
END; //
DELIMITER ;
Do the same for each of the other attribute tables.
Note that the id values are now unique across all your attribute tables.
Next you make members:abstract the target for a foreign key in members_history.
CREATE TABLE IF NOT EXISTS `members_history` (
`id` INT unsigned NOT NULL auto_increment COMMENT 'Unique Id',
`mid` INT unsigned NOT NULL COMMENT 'Members Unique Id.',
`value` INT UNSIGNED NOT NULL,
`table` enum('class','elected','appointed','status') NOT NULL,
`start` date NOT NULL COMMENT 'Value''s Effect Date',
`end` date default NULL COMMENT 'Value''s Expiration Date',
PRIMARY KEY (`id`),
FOREIGN KEY (`mid`) REFERENCES `members` (`id`) ON DELETE CASCADE,
FOREIGN KEY (`value`, `table`) REFERENCES `members:abstract` (`id`, `type`) ON DELETE CASCADE
) ENGINE=InnoDB DEFAULT CHARSET=latin1 COMMENT='Member History';
Notice that this table defines a foreign key so that you can't reference an id of the wrong type of attribute in members:abstract.
Now you can rely on referential integrity and consistency which is impossible when you try to implement polymorphic associations without the common parent of all the referenced attribute tables.
Here's the query that returns the result you described (tested on MySQL 5.1.40):
SELECT m.username,
m.password, m.salt, m.name_first, m.name_last,
MAX(a.name_prefix) AS name_prefix,
COALESCE(MAX(a.hours_extra), MAX(e.hours_extra)) AS hours_extra,
MAX(m.date_join) AS date_join,
MAX(m.date_leave) AS date_leave,
MAX(a.position) AS appointed,
MAX(c.class) AS class,
MAX(e.position) AS elected,
MAX(s.status) AS status
FROM `members` m
JOIN `members_history` h ON (h.mid = m.id)
LEFT OUTER JOIN `members:appointed` a ON (h.table = 'appointed' AND h.value = a.id)
LEFT OUTER JOIN `members:class` c ON (h.table = 'class' AND h.value = c.id)
LEFT OUTER JOIN `members:elected` e ON (h.table = 'elected' AND h.value = e.id)
LEFT OUTER JOIN `members:status` s ON (h.table = 'status' AND h.value = s.id)
GROUP BY m.id;
all you need is a left outer join for each of the history types and whatever logic you need to pick the "current" row.
your table structure doesn't quite make enough sense to me to put together a sample for you. maybe if you provide ONE sample member and a couple of rows of history for ONE attribute, i can help you out.