I am writing a script for my custom forums that determines if new threads or replies have been posted in a certain board since the user's last visit to that board(board_marks). I have two tables: one that stores threads, another that stores replies. I can easily find if there are any new replies in a board by using a left join. I want to add a bit that finds if there are new threads in that board. How would I do this?
My Tables:
CREATE TABLE IF NOT EXISTS `forum_threads` (
`thread_id` int(15) NOT NULL AUTO_INCREMENT,
`board_id` int(15) NOT NULL,
`author_id` int(15) NOT NULL,
`updater_id` int(15) NOT NULL,
`title` text NOT NULL,
`content` text NOT NULL,
`date_posted` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
`date_updated` timestamp NOT NULL DEFAULT '0000-00-00 00:00:00',
`views` int(15) NOT NULL,
`status` tinyint(1) NOT NULL,
`type` tinyint(1) NOT NULL COMMENT '0 normal, 1 sticky, 2 global.',
PRIMARY KEY (`thread_id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1;
CREATE TABLE IF NOT EXISTS `forum_messages` (
`message_id` int(15) NOT NULL AUTO_INCREMENT,
`thread_id` int(15) NOT NULL,
`author_id` int(15) NOT NULL,
`modifier_id` int(15) DEFAULT NULL,
`content` text NOT NULL,
`date_posted` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
`date_modified` timestamp NULL DEFAULT NULL,
`status` tinyint(1) NOT NULL,
PRIMARY KEY (`message_id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1;
My Script:
$this->db->select('m.message_id, m.thread_id, m.date_posted, t.thread_id, t.board_id');
$this->db->from('forum_messages AS m');
$this->db->join('forum_threads AS t', 'm.thread_id = t.thread_id', 'left');
$this->db->where('t.board_id', $board_id);
$this->db->where('m.date_posted > "'.$last_mark['date_marked'].'"');
if($query_new_messages = $this->db->get()){
if($query_new_messages->num_rows() > 0){
$contains_new_posts = TRUE;
} else {
$contains_new_posts = FALSE;
}
}
The actual query:
SELECT
m.message_id, m.thread_id, m.date_posted, t.thread_id, t.board_id
FROM (forum_messages AS m)
LEFT JOIN forum_threads AS t ON m.thread_id = t.thread_id
WHERE `t`.`board_id` = '10' AND `m`.`date_posted` > "2014-03-02 06:01:31"
PS: I am doing this in CodeIgniter.
An OR condition will get you the data you want:
SELECT
m.message_id, m.thread_id, m.date_posted, t.thread_id, t.board_id
FROM (forum_messages AS m)
LEFT JOIN forum_threads AS t ON m.thread_id = t.thread_id
WHERE `t`.`board_id` = '10'
AND (`m`.`date_posted` > "2014-03-02 06:01:31"
OR `t`.`date_posted` > "2014-03-02 06:01:31")
You should benchmark this, however. This solution is probably not very optimized. You should benchmark doing two fast queries - one for posts and one for threads - vs doing one potentially slower one.
Related
I have two tables messages and members. I tried joining tables without having a nested query but it does not reflect the join on members. So, I initially thought that I could do the following
SELECT M1.*, COUNT(M2.emid) AS replies FROM messages M1
LEFT JOIN messages M2
ON M2.thread = M1.emid
INNER JOIN members M
ON M.meid = M1.emitter
WHERE
M1.thread is NULL AND
M1.receiver = 2
GROUP BY
M1.emid
but it does not seem to join the corresponding member. Then I tried this and it gives me the result that I need but I would like to know if there is a way to accomplish the same result using joins without the nested query
SELECT * FROM (
SELECT M1.*, COUNT(M2.emid) AS replies FROM messages M1
LEFT JOIN messages M2
ON M2.thread = M1.emid
WHERE
M1.thread is NULL AND
M1.receiver = 2
GROUP BY
M1.emid
) O INNER JOIN members M ON O.receiver = M.meid
-- Table structure for table members
CREATE TABLE `members` (
`meid` bigint(64) NOT NULL,
`name` varchar(32) DEFAULT NULL,
`lastname` varchar(32) DEFAULT NULL,
`email` varchar(128) NOT NULL,
`mobile` char(10) DEFAULT NULL,
`college` bigint(64) NOT NULL,
`major` bigint(64) NOT NULL,
`password` varchar(256) NOT NULL,
`oauth` varchar(128) DEFAULT NULL,
`confirmed` tinyint(4) DEFAULT NULL,
`active` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
`joined` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP
) ENGINE=InnoDB DEFAULT CHARSET=latin1;
-- Table structure for table messages
CREATE TABLE `messages` (
`emid` bigint(20) NOT NULL,
`emitter` bigint(20) NOT NULL,
`receiver` bigint(20) NOT NULL,
`thread` bigint(20) DEFAULT NULL,
`opened` tinyint(4) DEFAULT '0',
`message` blob NOT NULL,
`timecard` datetime DEFAULT CURRENT_TIMESTAMP
) ENGINE=InnoDB DEFAULT CHARSET=latin1;
trying to query a large table (senddb.order_histories) that has close to 50M rows and this is the MySQL query I am using:
FIRST APPROACH- inner join:
select a.id,
a.order_number,
a.sku_id,
a.fulfillment_status,
a.modified_by,
a.created_at,
a.updated_at
from senddb.order_line_items a
inner join (
select order_line_item_id,
order_number,
order_status,
order_status_description,
action,
modified_by,
created_at,
max(updated_at) as updated_at
from senddb.order_histories
where order_status in ('x','y','z')
and fulfillment_location = 'abcd'
group by order_line_item_id) as b
on a.id = b.order_line_item_id
and a.fulfillment_status = '2';
EXPLAIN output :
SECOND APPROACH- nested select:
select a.id,
a.order_number,
a.sku_id,
a.fulfillment_status,
a.modified_by,
a.created_at,
a.updated_at
from senddb.order_line_items a
where a.fulfillment_status = '2'
and a.id in (
select b.order_line_item_id from(
select order_line_item_id,
order_number,
order_status,
order_status_description,
action,
modified_by,
created_at,
max(updated_at) as updated_at
from senddb.order_histories
where
order_status in ('x','y','z')
and fulfillment_location = 'abcd'
group by order_line_item_id) as b);
I believe nested select is a bad approach on large data but i anyhow added it here because it worked on my sample set. Anyway both the queries eventually time out after 600 seconds with the message : Error Code: 2013. Lost connection to MySQL server during query.
I would like to know if there are any ways to alter the query to make it run faster. I have already tried reducing the columns in the inner select / inner join but that should not really be an issue IMO. I also looked up a solution that says "create a clustered index" but i wasn't really able to follow. Any help is appreciated.
TABLE order_histories :
order_histories CREATE TABLE `order_histories` (
`id` int(4) unsigned NOT NULL AUTO_INCREMENT,
`order_number` varchar(24) DEFAULT NULL,
`order_status_description` varchar(255) DEFAULT NULL,
`datetime_stamp` datetime DEFAULT NULL,
`action` varchar(32) DEFAULT NULL,
`fulfillment_location` int(8) DEFAULT NULL,
`order_status` int(8) DEFAULT NULL,
`user_id` int(8) DEFAULT NULL,
`created_at` datetime DEFAULT NULL,
`updated_at` datetime DEFAULT NULL,
`modified_by` varchar(32) DEFAULT NULL,
`order_line_item_id` int(11) DEFAULT NULL,
`pooled` tinyint(1) DEFAULT '0',
PRIMARY KEY (`id`),
KEY `order_histories_ecash_idx` (`order_number`),
KEY `order_line_item_id` (`order_line_item_id`)
) ENGINE=InnoDB AUTO_INCREMENT=454738178 DEFAULT CHARSET=latin1
TABLE order_line_items :
order_line_items CREATE TABLE `order_line_items` (
`id` int(4) unsigned NOT NULL AUTO_INCREMENT,
`order_number` varchar(24) DEFAULT NULL,
`sku_id` int(8) DEFAULT NULL,
`original_price` float DEFAULT NULL,
`dept_description` varchar(100) DEFAULT NULL,
`description` varchar(100) DEFAULT NULL,
`quantity_ordered` int(8) DEFAULT NULL,
`gift_indicator` char(1) DEFAULT NULL,
`gift_wrap_flag` char(1) DEFAULT NULL,
`shipping_record_flag` char(1) DEFAULT NULL,
`gift_comments` varchar(100) DEFAULT NULL,
`item_status` char(1) DEFAULT NULL,
`tax_amount` float DEFAULT NULL,
`tax_rate` float DEFAULT NULL,
`upc` varchar(20) DEFAULT NULL,
`final_price` float DEFAULT NULL,
`line_number` int(8) DEFAULT NULL,
`master_line_number` int(8) DEFAULT NULL,
`gift_wrap_flag_type` char(1) DEFAULT NULL,
`color_code` varchar(4) DEFAULT NULL,
`size_id` varchar(6) DEFAULT NULL,
`width_id` varchar(6) DEFAULT NULL,
`brand` varchar(15) DEFAULT NULL,
`vpn` varchar(30) DEFAULT NULL,
`dept_number` int(8) DEFAULT NULL,
`class_number` int(8) DEFAULT NULL,
`non_merch_item` char(1) DEFAULT NULL,
`created_at` datetime DEFAULT NULL,
`updated_at` datetime DEFAULT NULL,
`modified_by` varchar(32) DEFAULT NULL,
`chain_id` int(11) DEFAULT NULL,
`fulfillment_location` int(11) DEFAULT NULL,
`fulfillment_date` datetime DEFAULT NULL,
`fulfillment_status` int(11) DEFAULT NULL,
`fulfillment_sales_associate` int(11) DEFAULT NULL,
`gift_wrap_line_number` int(11) DEFAULT NULL,
`shipping_type` int(11) DEFAULT NULL,
`order_track_info_id` int(11) DEFAULT NULL,
`store_tlog_updated` varchar(1) DEFAULT NULL,
`shipping_tlx_code` int(11) DEFAULT NULL,
`store_closed` tinyint(1) DEFAULT NULL,
`flags` int(11) DEFAULT NULL,
`deal_based_index` int(11) DEFAULT NULL,
`tlog_calc_ret_price` float DEFAULT NULL,
`tlog_amount` float DEFAULT NULL,
`tlog_retail_price` float DEFAULT NULL,
`tlog_ext_amount` float DEFAULT NULL,
`tlog_flag_1` int(11) DEFAULT NULL,
`tlog_flag_2` int(11) DEFAULT NULL,
`tlog_flag_3` int(11) DEFAULT NULL,
`time_remaining` int(11) DEFAULT NULL,
PRIMARY KEY (`id`),
KEY `order_line_items_ecash_idx` (`order_number`),
KEY `order_line_item_fulfillment_location_idx` (`fulfillment_location`),
KEY `order_line_item_fulfillment_status_idx` (`fulfillment_status`),
KEY `upc_idx` (`upc`),
KEY `sku_id_idx` (`sku_id`),
KEY `order_line_items_idx001` (`order_number`,`id`,`fulfillment_status`),
KEY `order_track_info_id` (`order_track_info_id`),
KEY `shipping_type_idx` (`shipping_type`,`non_merch_item`) USING BTREE
) ENGINE=InnoDB AUTO_INCREMENT=11367052 DEFAULT CHARSET=latin1
This query can be simplified:
select a.id,
a.order_number,
a.sku_id,
a.fulfillment_status,
a.modified_by,
a.created_at,
a.updated_at
from senddb.order_line_items a
inner join senddb.order_histories b on a.id = b.order_line_item_id
where b.order_status in ('x','y','z')
and b.fulfillment_location = 'abcd'
and a.fulfillment_status = '2';
Since you're only selecting values from table a, you don't need to select specific values from table b and can instead just apply your conditions. Outside of this, you need to ensure that b.order_line_item_id has an index on it. You can find more about creating indexes here. I'm not an expert in MySQL but something similar to this should work if senddb.order_histories.order_line_item_id isn't already the primary key.
CREATE INDEX IX_order_histories_order_line_item_id ON order_histories (order_line_item_id);
You need to read up the optimization section of the MySQL docs. It contains a lot of information on how you can optimize your queries and data sets. The main idea here is to add indexes to the fields that are being used as the criteria in the WHERE clause of the SQL statements.
Basically, both of your alternatives are using a "sub-SELECT, not an INNER JOIN.
The syntax of a true JOIN is one of the following:
SELECT ...
FROM X INNER JOIN Y USING (field_list)
... or ...
SELECT ...
FROM X INNER JOIN Y ON (x.field1 = y.field2) ...
But in both cases the objects being joined are tables or views.
I'm going to presume ... admittedly, without checking ... that Nick Larsen's answer #1 adequately re-expresses your original query using JOINs.
(Notice how, in his answer, the shorthand identifiers A and B are introduced as referring to each of the two table-names mentioned in his query.)
Firstly, you need to decide if a 50 million resultset is what you are asking for. Big data tables are not there so that you can select all their rows. They are there so that you can ask them questions using sql queries. SQL is a query language, it's not a data loading language.
What's your purpose? If you want to copy the data you can do that by loading the data, for example, 1000 rows per query in a for loop. if you are loading the data for processing, you can do that in the same way.
If you want to derive statistical information, you can use outer join and return a low number of rows, using aggregate functions. But you shouldn't do that either, what you "should" do is to decide what you want from the table and preferably, run aggregate functions to store useful information in a different table. (mostly SELECT INTO queries) You should never need to join a table of 50 million records in the first place.
Telling you how to do something wrong using indexes wouldn't be the right thing here.
I have an sql query that collects the year to date totals for all products in the inventory. This query runs quickly, returning around 5000 results in a little over a second.
The query used is
SELECT `ITINCODE` as Code, `IT_product_id` as ID,
SUM(IF(YEAR(`ITDATE`) = YEAR(CURDATE() ), IF(ITTYPE = "O",`ITVAL`, -`ITVAL` ), 0) ) as 'YTD_VAL'
FROM `FITEMS`
WHERE (ITTYPE = 'O' OR `ITTYPE` = 'R') AND ITBACKORDER = 'N' AND ITAMNT > 0 AND YEAR(`ITDATE`) >= YEAR(CURDATE() ) -1
GROUP BY `ITINCODE`
ORDER BY YTD_VAL DESC
I want to take the values in YTD_VAL and store them in the actual products table so that they can be used as an ORDER BY on other queries. I added a new field called ytd_val to the products table and then ran
UPDATE products p SET p.ytd_val =
(SELECT SUM(IF(YEAR(`ITDATE`) = YEAR(CURDATE() ), IF(ITTYPE = 'O',`ITVAL`, -`ITVAL` ), 0) ) as 'YTD_VAL'
FROM `FITEMS`
WHERE ITINCODE = p.products_model AND (ITTYPE = 'O' OR `ITTYPE` = 'R') AND ITBACKORDER = 'N' AND ITAMNT > 0 AND YEAR(`ITDATE`) >= YEAR(CURDATE() ) -1
GROUP BY `ITINCODE`
ORDER BY YTD_VAL DESC)
The idea was to run this by cron job each night so that the values were updated to reflect the previous days sales.
However, running this query has taken 10 plus minutes and it still hasn't completed.
ITINCODE in FITEMS table is the same as products_model in the products table.
IT_product_id in the FITEMS table is the same as products_id in the products table.
What can I do to speed up the query? I thought that as the original results query returns quickly enough that simply updating the values on another table would take seconds longer!
Table structure is as follows:
show create table fitems\G;
Create Table: CREATE TABLE `fitems` (
`ITUNIQUEREF` int(11) unsigned NOT NULL,
`ITAMNT` int(11) NOT NULL,
`ITREF` int(11) unsigned NOT NULL,
`ITTYPE` char(1) NOT NULL,
`ITVAL` decimal(10,4) NOT NULL,
`ITVAT` decimal(10,4) NOT NULL,
`ITPRICE` decimal(10,4) NOT NULL,
`ITDATE` date NOT NULL,
`ITBACKORDER` char(1) NOT NULL,
`ITDISC` decimal(10,2) NOT NULL,
`ITORDERREF` int(11) NOT NULL,
`ITTREF` int(11) unsigned NOT NULL,
`ITDATEDLY` date NOT NULL,
`ITINCODE` char(20) NOT NULL,
`IT_product_id` int(11) unsigned NOT NULL,
`ITBUILT` int(11) NOT NULL,
PRIMARY KEY (`ITUNIQUEREF`),
KEY `ITREF` (`ITREF`,`ITTYPE`,`ITDATE`,`ITBACKORDER`,`ITORDERREF`,`ITDATEDLY`,`ITINCODE`,`IT_product_id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1;
.
show create table products\G;
CREATE TABLE `products` (
`products_id` int(11) NOT NULL,
`products_type` int(11) NOT NULL DEFAULT '1',
`products_quantity` float NOT NULL DEFAULT '0',
`products_model` varchar(32) CHARACTER SET utf8 DEFAULT NULL,
`products_image` varchar(64) CHARACTER SET utf8 DEFAULT NULL,
`products_price` decimal(15,4) NOT NULL DEFAULT '0.0000',
`products_group_a_price` decimal(15,4) NOT NULL,
`products_group_b_price` decimal(15,4) NOT NULL,
`products_group_c_price` decimal(15,4) NOT NULL,
`products_group_d_price` decimal(15,4) NOT NULL,
`products_group_e_price` decimal(15,4) NOT NULL,
`products_group_f_price` decimal(15,4) NOT NULL,
`products_group_g_price` decimal(15,4) NOT NULL,
`products_virtual` tinyint(1) NOT NULL DEFAULT '0',
`products_date_added` datetime NOT NULL DEFAULT '0001-01-01 00:00:00',
`products_last_modified` datetime DEFAULT NULL,
`products_date_available` datetime DEFAULT NULL,
`products_weight` float NOT NULL DEFAULT '0',
`products_status` tinyint(1) NOT NULL DEFAULT '0',
`products_tax_class_id` int(11) NOT NULL DEFAULT '0',
`manufacturers_id` int(11) DEFAULT NULL,
`products_ordered` float NOT NULL DEFAULT '0',
`products_quantity_order_min` float NOT NULL DEFAULT '1',
`products_quantity_order_units` float NOT NULL DEFAULT '1',
`products_priced_by_attribute` tinyint(1) NOT NULL DEFAULT '0',
`product_is_free` tinyint(1) NOT NULL DEFAULT '0',
`product_is_call` tinyint(1) NOT NULL DEFAULT '0',
`products_quantity_mixed` tinyint(1) NOT NULL DEFAULT '0',
`product_is_always_free_shipping` tinyint(1) NOT NULL DEFAULT '0',
`products_qty_box_status` tinyint(1) NOT NULL DEFAULT '1',
`products_quantity_order_max` float NOT NULL DEFAULT '0',
`products_sort_order` int(11) NOT NULL DEFAULT '0',
`products_canonical` text COLLATE utf8_unicode_ci NOT NULL,
`products_discount_type` tinyint(1) NOT NULL DEFAULT '0',
`products_discount_type_from` tinyint(1) NOT NULL DEFAULT '0',
`products_price_sorter` decimal(15,4) NOT NULL DEFAULT '0.0000',
`master_categories_id` int(11) NOT NULL DEFAULT '0',
`products_mixed_discount_quantity` tinyint(1) NOT NULL DEFAULT '1',
`metatags_title_status` tinyint(1) NOT NULL DEFAULT '0',
`metatags_products_name_status` tinyint(1) NOT NULL DEFAULT '0',
`metatags_model_status` tinyint(1) NOT NULL DEFAULT '0',
`metatags_price_status` tinyint(1) NOT NULL DEFAULT '0',
`metatags_title_tagline_status` tinyint(1) NOT NULL DEFAULT '0',
`pricing_group` varchar(16) COLLATE utf8_unicode_ci DEFAULT NULL,
`ytd_val` int(20) NOT NULL,
PRIMARY KEY (`products_id`),
KEY `idx_products_date_added_zen` (`products_date_added`),
KEY `idx_products_status_zen` (`products_status`),
KEY `idx_products_date_available_zen` (`products_date_available`),
KEY `idx_products_ordered_zen` (`products_ordered`),
KEY `idx_products_model_zen` (`products_model`),
KEY `idx_products_price_sorter_zen` (`products_price_sorter`),
KEY `idx_master_categories_id_zen` (`master_categories_id`),
KEY `idx_products_sort_order_zen` (`products_sort_order`),
KEY `idx_manufacturers_id_zen` (`manufacturers_id`),
KEY `products_price` (`products_price`),
KEY `products_status_products_price` (`products_status`,`products_price`),
FULLTEXT KEY `idx_enhanced_products_model` (`products_model`)
) ENGINE=MyISAM DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci
SELECT is always way faster then UPDATE.
To speed up an update:
Look at the indexes on the table you are updating: Are they all needed? If not, remove unneeded ones (I would at least remove idx_products_status_zen since that is also covered by products_status_products_price)
Look at the data model: Can you partition the table you are updating? If so, that will speed up the update since the indexes you are updating will be smaller and thus quicker to update;
Use InnoDB. It is faster;
Do you need to be ACID compliant? If not, then change the settings of InnoDB to speed up the system.
If you are really using MySQL: Switch to MariaDB: It is about 8% faster;
Install monitoring to see where your bottleneck is: IO or CPU: If it is IO on read, then try compression.
If you are satisfied with first query performance.
You can just transform your second query to use INNER JOIN like:
http://sqlfiddle.com/#!9/1028a/1
UPDATE products p
INNER JOIN (
SELECT SUM(IF(YEAR(`ITDATE`) = YEAR(CURDATE() ), IF(ITTYPE = 'O',`ITVAL`, -`ITVAL` ), 0) ) as 'YTD_VAL',
ITINCODE
FROM `FITEMS`
WHERE (ITTYPE = 'O' OR `ITTYPE` = 'R')
AND ITBACKORDER = 'N' AND ITAMNT > 0 AND YEAR(`ITDATE`) >= YEAR(CURDATE() ) -1
GROUP BY `ITINCODE`
) t
ON t.ITINCODE = p.products_model
SET p.ytd_val = t.YTD_VAL
UPDATE And by the way you don't need that ORDER BY YTD_VAL DESC it has no sense in this particular case I guess.
UPDATE 2
UPDATE products p
INNER JOIN (
SELECT SUM(IF(YEAR(`ITDATE`) = YEAR(CURDATE() ), IF(ITTYPE = 'O',`ITVAL`, -`ITVAL` ), 0) ) as 'YTD_VAL',
IT_product_id
FROM `FITEMS`
WHERE (ITTYPE = 'O' OR `ITTYPE` = 'R')
AND ITBACKORDER = 'N' AND ITAMNT > 0 AND YEAR(`ITDATE`) >= YEAR(CURDATE() ) -1
GROUP BY `IT_product_id`
) t
ON t.IT_product_id = p.products_id
SET p.ytd_val = t.YTD_VAL
I am not one to say do it, but consider the case for Covering Indexes. One in particular on table fitems. Those columns are relatively slim for that query. I will spook up one to try on particular columns. We have seen cases where terribly long queries can be completed in snappy time. But no promises. Ah, I see you have some. Was looking at the first edit with the alter tables below it. I will keep looking.
Covering Index
A covering index is one in which the query can be resolved via a quick stroll through the index b-tree, without requiring a lookup into the actual table data page. These are the ultimate nirvana for speed. Percona quick article on it.
Explain
And run the query (the update one) thru Explain and examine the output. See also the article Using Explain to Write Better Mysql Queries.
Note, some of these comments are for those that follows, not necessarily this op. He seems to have his house in order pretty well.
In your subquery, neither the order by nor the group by are necessary, so the update can be written as:
UPDATE products p
SET p.ytd_val = (SELECT SUM(IF(YEAR(`ITDATE`) = YEAR(CURDATE() ), IF(ITTYPE = 'O',`ITVAL`, -`ITVAL` ), 0) )
FROM `FITEMS`
WHERE FITEMS.ITINCODE = p.products_model AND
ITTYPE IN ('O', 'R') AND
ITBACKORDER = 'N' AND
ITAMNT > 0 AND
YEAR(`ITDATE`) >= YEAR(CURDATE() ) - 1
);
For this, you want an index on FITEMS(ITINCODE, ITBACKORDER, ITTYPE, ITAMNT, ITVAL). This might significantly speed your query.
I have three tables:
CREATE TABLE IF NOT EXISTS `contacts` (
`id` int(10) unsigned NOT NULL AUTO_INCREMENT,
`providerId` int(10) unsigned NOT NULL DEFAULT '0',
`requestId` int(10) unsigned NOT NULL DEFAULT '0',
`status` binary(1) NOT NULL DEFAULT '0',
PRIMARY KEY (`id`)
)
CREATE TABLE IF NOT EXISTS `messages` (
`id` int(255) NOT NULL AUTO_INCREMENT,
`fromuid` int(255) NOT NULL,
`touid` int(255) NOT NULL,
`sentdt` datetime NOT NULL,
`read` tinyint(1) NOT NULL DEFAULT '0',
`readdt` datetime DEFAULT NULL,
`messagetext` longtext CHARACTER SET utf8 COLLATE utf8_bin NOT NULL,
PRIMARY KEY (`id`),
KEY `id` (`id`)
)
CREATE TABLE IF NOT EXISTS `users` (
`id` bigint(11) NOT NULL AUTO_INCREMENT,
`name` varchar(255) NOT NULL,
`email` varchar(255) DEFAULT NULL,
`mobile` varchar(15) NOT NULL,
`password` varchar(255) NOT NULL,
`city` varchar(255) NOT NULL,
`zip` varchar(15) DEFAULT NULL,
`device` varchar(50) DEFAULT NULL,
`version` varchar(10) DEFAULT NULL,
`photo` varchar(255) DEFAULT NULL,
`created` datetime NOT NULL,
`live` enum('0','1') NOT NULL DEFAULT '1',
`authenticationTime` datetime NOT NULL,
`userKey` varchar(255) DEFAULT NULL,
`IP` varchar(50) DEFAULT NULL,
`port` int(10) DEFAULT '0',
PRIMARY KEY (`id`),
KEY `firstname` (`mobile`,`city`,`zip`)
)
And this SQL query that finds out friends/contacts for specified user (user id 1 in this case):
SELECT u.id
,u.mobile
,u.name
,(NOW() - u.authenticationTime) AS authenticateTimeDifference
,u.IP
,f.providerid
,f.requestid
,f.status
,u.port
FROM contacts f
LEFT JOIN users u ON u.id =
IF (
f.providerid = 1
,f.requestid
,f.providerid
) WHERE (
f.providerid = 1
AND f.status = 1
)
OR f.requestid = 1
That works fine but I want to be able to also join messages table and show user's friends/contacts who have talked latest (meaning latest conversations first) with order by messages.sentdt desc option but I am unable to figure out how to do that, I tried all joins but none worked :(
Your help will be greatly appreciated. Thanks
Update
Here is sample data above query returns:
In that same resultset, I want to be able to sort based on order by messages.sentdt desc but I am not sure how to pull that in and sort resultset by latest message first
Try this:
select u.id
, u.mobile
, u.name
, (NOW() - u.authenticationTime) as authenticateTimeDifference
, u.IP
, f.providerid
, f.requestid
, f.status
, u.port
from contacts f
left join users u
on u.id = if (f.providerid = 1, f.requestid, f.providerid)
left join (select fromuid, max(sentdt) as sentdt from messages group by fromuid) m
on m.fromuid = if (f.providerid = 1, f.providerid, f.requestid)
where (f.providerid = 1 and f.status = 1)
or f.requestid = 1
order by m.sentdt
I have two MySQL tables: tech_requests and comments. I want to display each tech_request one time in a list ordered by the "last modified" date, whether that be the date of the tech_request creation or the latest comment tied to that tech_request. I was trying to use UNION but I got stuck. Any ideas would be much appreciated. Here are the tables:
CREATE TABLE `tech_requests` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`staff_member_id` int(3) NOT NULL,
`date_time` datetime NOT NULL,
`request` text NOT NULL,
`building_id` int(2) NOT NULL,
`technician_id` int(2) DEFAULT NULL,
`completed` tinyint(1) NOT NULL,
`subject` varchar(30) NOT NULL DEFAULT '',
`category_id` int(2) NOT NULL,
PRIMARY KEY (`id`)
) ENGINE=InnoDB AUTO_INCREMENT=203 DEFAULT CHARSET=utf8;
CREATE TABLE `comments` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`tech_request_id` int(11) NOT NULL,
`technician_id` int(2) NOT NULL,
`date_time` datetime NOT NULL,
`comment` text NOT NULL,
PRIMARY KEY (`id`)
) ENGINE=InnoDB AUTO_INCREMENT=234 DEFAULT CHARSET=utf8;
Are you looking for something like this?
SELECT r.id, r.staff_member_id, ...,
GREATEST(r.date_time, COALESCE(c.date_time, 0)) last_modified
FROM tech_requests r LEFT JOIN
(
SELECT tech_request_id, MAX(date_time) date_time
FROM comments c
GROUP BY tech_request_id
) c
ON r.id = c.tech_request_id
ORDER BY last_modified
Here is SQLFiddle demo