CODEWITHSUNDEEP
mysql
Beetlegeuse
Beetlegeuse

Reputation: 297

Optimizing query mysql

I have 3 tables in which i store some values. Everything is working fine except my query is taking too long to execute. I have around 500.000 rows in table "tickets" at the moment. I would like to know what would be the best way to optimize this query to execute SELECT faster.

One more thing: I would like to know is there a way to update all rows through query (not using my C# app). In this case i need to update column "wonamount" which is in tickets with value that i get by multiplying row "coefficient" with row "uplata" and update column status with value "2".

Here are my tables and sql:

CREATE TABLE IF NOT EXISTS `coefficients` (`number` int(11) DEFAULT NULL,`coefficient` int(11) DEFAULT NULL) ENGINE=InnoDB DEFAULT CHARSET=latin1;



INSERT INTO `coefficients` (`number`, `coefficient`) VALUES
(1, 0),
(2, 0),
(3, 0),
(4, 0),
(5, 0),
(6, 10000),
(7, 7500),
(8, 5000),
(9, 2500),
(10, 1000),
(11, 500),
(12, 300),
(13, 200),
(14, 120),
(15, 80),
(16, 70),
(17, 60),
(18, 50),
(19, 40),
(20, 35),
(21, 30),
(22, 25),
(23, 20),
(24, 15),
(25, 12),
(26, 10),
(27, 9),
(28, 8),
(29, 7),
(30, 6),
(31, 5),
(32, 4),
(33, 3),
(34, 2),
(35, 1);


CREATE TABLE IF NOT EXISTS `draws` (
`iddraws` int(11) NOT NULL AUTO_INCREMENT,
`1` varchar(255) DEFAULT NULL,
`2` varchar(45) DEFAULT NULL,
`3` varchar(45) DEFAULT NULL,
`4` varchar(45) DEFAULT NULL,
`5` varchar(45) DEFAULT NULL,
`6` varchar(45) DEFAULT NULL,
`7` varchar(45) DEFAULT NULL,
`8` varchar(45) DEFAULT NULL,
`9` varchar(45) DEFAULT NULL,
`10` varchar(45) DEFAULT NULL,
`11` varchar(45) DEFAULT NULL,
`12` varchar(45) DEFAULT NULL,
`13` varchar(45) DEFAULT NULL,
`14` varchar(45) DEFAULT NULL,
`15` varchar(45) DEFAULT NULL,
`16` varchar(45) DEFAULT NULL,
`17` varchar(45) DEFAULT NULL,
`18` varchar(45) DEFAULT NULL,
`19` varchar(45) DEFAULT NULL,
`20` varchar(45) DEFAULT NULL,
`21` varchar(45) DEFAULT NULL,
`22` varchar(45) DEFAULT NULL,
`23` varchar(45) DEFAULT NULL,
`24` varchar(45) DEFAULT NULL,
`25` varchar(45) DEFAULT NULL,
`26` varchar(45) DEFAULT NULL,
`27` varchar(45) DEFAULT NULL,
`28` varchar(45) DEFAULT NULL,
`29` varchar(45) DEFAULT NULL,
`30` varchar(45) DEFAULT NULL,
`31` varchar(45) DEFAULT NULL,
`32` varchar(45) DEFAULT NULL,
`33` varchar(45) DEFAULT NULL,
`34` varchar(45) DEFAULT NULL,
`35` varchar(45) DEFAULT NULL,
`datetime` varchar(255) DEFAULT NULL,
PRIMARY KEY (`iddraws`)
) ENGINE=InnoDB  DEFAULT CHARSET=latin1 AUTO_INCREMENT=162 ;


INSERT INTO `draws` (`iddraws`, `1`, `2`, `3`, `4`, `5`, `6`, `7`, `8`, `9`, `10`, `11`, `12`, `13`, `14`, `15`, `16`, `17`, `18`, `19`, `20`, `21`, `22`, `23`, `24`, `25`, `26`, `27`, `28`, `29`, `30`, `31`, `32`, `33`, `34`, `35`, `datetime`) VALUES
(1, '17', '46', '27', '30', '8', '11', '4', '40', '37', '36', '22', '14', '35', '47', '24', '20', '23', '10', '2', '42', '41', '43', '9', '19', '7', '48', '3', '38', '29', '44', '16', '12', '26', '13', '5', '1391130262'),
(2, '45', '2', '1', '24', '30', '4', '10', '11', '22', '3', '38', '33', '35', '14', '48', '28', '42', '27', '43', '9', '15', '29', '36', '41', '26', '23', '13', '5', '16', '20', '12', '6', '32', '37', '19', '1391134904'),
(3, '12', '46', '32', '15', '14', '41', '45', '6', '9', '20', '26', '2', '47', '37', '33', '39', '34', '17', '16', '23', '35', '29', '44', '36', '18', '40', '22', '4', '27', '30', '38', '21', '3', '43', '24', '1391135221');

CREATE TABLE IF NOT EXISTS `tickets` (
`id_tiketa` int(11) NOT NULL AUTO_INCREMENT,
`idtickets` varchar(45) DEFAULT NULL,
`b1` varchar(45) DEFAULT NULL,
`b2` varchar(45) DEFAULT NULL,
`b3` varchar(45) DEFAULT NULL,
`b4` varchar(45) DEFAULT NULL,
`b5` varchar(45) DEFAULT NULL,
`b6` varchar(45) DEFAULT NULL,
`shop` varchar(45) DEFAULT NULL,
`user` varchar(45) DEFAULT NULL,
`time` varchar(255) DEFAULT NULL,
`status` varchar(45) DEFAULT NULL,
`uplata` varchar(45) DEFAULT NULL,
`draw` varchar(45) DEFAULT NULL,
`qt` varchar(45) DEFAULT NULL,
`wonamount` varchar(45) DEFAULT NULL,
PRIMARY KEY (`id_tiketa`)
) ENGINE=InnoDB  DEFAULT CHARSET=latin1 AUTO_INCREMENT=138 ;


INSERT INTO `tickets` (`id_tiketa`, `idtickets`, `b1`, `b2`, `b3`, `b4`, `b5`, `b6`, `shop`, `user`, `time`, `status`, `uplata`, `draw`, `qt`, `wonamount`) VALUES
(75, '4-1-170-1367', '41', '47', '17', '24', '15', '44', '170', 'w1', '1391149398', '1', '1', '1', '', ''),
(76, '4-1-170-20104', '23', '27', '13', '7', '14', '42', '170', 'w1', '1391149398', '1', '1', '1', '', ''),
(91, '4-2-170-13887', '16', '4', '13', '35', '30', '9', '170', 'w1', '1391149462', '1', '1', '2', '', ''),
(92, '4-2-170-9701', '2', '32', '7', '15', '5', '34', '170', 'w1', '1391149463', '1', '1', '2', '', ''),
(93, '4-2-170-45661', '23', '24', '22', '27', '48', '6', '170', 'w1', '1391149463', '1', '1', '2', '', ''),
(98, '4-2-170-45503', '36', '13', '33', '10', '29', '9', '170', 'w1', '1391149463', '1', '1', '2', '', ''),
(99, '4-2-170-24095', '19', '35', '11', '36', '46', '40', '170', 'w1', '1391149463', '1', '1', '2', '', ''),
(100, '4-2-170-42832', '27', '32', '17', '29', '7', '21', '170', 'w1', '1391149463', '1', '1', '2', '', ''),
(101, '4-2-170-13570', '22', '23', '32', '6', '1', '28', '170', 'w1', '1391149463', '1', '1', '2', '', ''),
(103, '4-2-170-28122', '15', '10', '11', '9', '14', '48', '170', 'w1', '1391149463', '1', '1', '2', '', ''),
(116, '4-2-170-13095', '28', '20', '33', '42', '26', '14', '170', 'w1', '1391149464', '1', '1', '2', '', ''),
(118, '4-2-170-27646', '23', '14', '37', '27', '24', '19', '170', 'w1', '1391149464', '1', '1', '2', '', ''),
(124, '4-2-170-23302', '20', '23', '15', '38', '4', '45', '170', 'w1', '1391149465', '1', '1', '2', '', '');







SELECT t.idtickets,
       t.uplata,
       c.coefficient
FROM tickets t
INNER JOIN draws d ON(FIELD(t.b1,d.1,d.2,d.3,d.4,d.5,d.6,d.7,d.8,d.9,d.10,d.11,d.12,d.13,d.14,d.15,d.16,d.17,d.18,d.19,d.20,d.21,d.22,d.23,d.24,d.25,d.26,d.27,d.28,d.29,d.30,d.31,d.32,d.33,d.34,d.35)>0)
AND (FIELD(t.b2,d.1,d.2,d.3,d.4,d.5,d.6,d.7,d.8,d.9,d.10,d.11,d.12,d.13,d.14,d.15,d.16,d.17,d.18,d.19,d.20,d.21,d.22,d.23,d.24,d.25,d.26,d.27,d.28,d.29,d.30,d.31,d.32,d.33,d.34,d.35)>0)
AND (FIELD(t.b3,d.1,d.2,d.3,d.4,d.5,d.6,d.7,d.8,d.9,d.10,d.11,d.12,d.13,d.14,d.15,d.16,d.17,d.18,d.19,d.20,d.21,d.22,d.23,d.24,d.25,d.26,d.27,d.28,d.29,d.30,d.31,d.32,d.33,d.34,d.35)>0)
AND (FIELD(t.b4,d.1,d.2,d.3,d.4,d.5,d.6,d.7,d.8,d.9,d.10,d.11,d.12,d.13,d.14,d.15,d.16,d.17,d.18,d.19,d.20,d.21,d.22,d.23,d.24,d.25,d.26,d.27,d.28,d.29,d.30,d.31,d.32,d.33,d.34,d.35)>0)
AND (FIELD(t.b5,d.1,d.2,d.3,d.4,d.5,d.6,d.7,d.8,d.9,d.10,d.11,d.12,d.13,d.14,d.15,d.16,d.17,d.18,d.19,d.20,d.21,d.22,d.23,d.24,d.25,d.26,d.27,d.28,d.29,d.30,d.31,d.32,d.33,d.34,d.35)>0)
AND (FIELD(t.b6,d.1,d.2,d.3,d.4,d.5,d.6,d.7,d.8,d.9,d.10,d.11,d.12,d.13,d.14,d.15,d.16,d.17,d.18,d.19,d.20,d.21,d.22,d.23,d.24,d.25,d.26,d.27,d.28,d.29,d.30,d.31,d.32,d.33,d.34,d.35)>0)
INNER JOIN coefficients c ON c.number = GREATEST(FIELD(t.b1,d.1,d.2,d.3,d.4,d.5,d.6,d.7,d.8,d.9,d.10,d.11,d.12,d.13,d.14,d.15,d.16,d.17,d.18,d.19,d.20,d.21,d.22,d.23,d.24,d.25,d.26,d.27,d.28,d.29,d.30,d.31,d.32,d.33,d.34,d.35), FIELD(t.b2,d.1,d.2,d.3,d.4,d.5,d.6,d.7,d.8,d.9,d.10,d.11,d.12,d.13,d.14,d.15,d.16,d.17,d.18,d.19,d.20,d.21,d.22,d.23,d.24,d.25,d.26,d.27,d.28,d.29,d.30,d.31,d.32,d.33,d.34,d.35), FIELD(t.b3,d.1,d.2,d.3,d.4,d.5,d.6,d.7,d.8,d.9,d.10,d.11,d.12,d.13,d.14,d.15,d.16,d.17,d.18,d.19,d.20,d.21,d.22,d.23,d.24,d.25,d.26,d.27,d.28,d.29,d.30,d.31,d.32,d.33,d.34,d.35), FIELD(t.b4,d.1,d.2,d.3,d.4,d.5,d.6,d.7,d.8,d.9,d.10,d.11,d.12,d.13,d.14,d.15,d.16,d.17,d.18,d.19,d.20,d.21,d.22,d.23,d.24,d.25,d.26,d.27,d.28,d.29,d.30,d.31,d.32,d.33,d.34,d.35), FIELD(t.b5,d.1,d.2,d.3,d.4,d.5,d.6,d.7,d.8,d.9,d.10,d.11,d.12,d.13,d.14,d.15,d.16,d.17,d.18,d.19,d.20,d.21,d.22,d.23,d.24,d.25,d.26,d.27,d.28,d.29,d.30,d.31,d.32,d.33,d.34,d.35), FIELD(t.b6,d.1,d.2,d.3,d.4,d.5,d.6,d.7,d.8,d.9,d.10,d.11,d.12,d.13,d.14,d.15,d.16,d.17,d.18,d.19,d.20,d.21,d.22,d.23,d.24,d.25,d.26,d.27,d.28,d.29,d.30,d.31,d.32,d.33,d.34,d.35))
WHERE t.draw='1'
  AND t.status = '1'
  AND d.iddraws='1'

And yes, i need to do that for each t.draw and d.iddraws which will be same values.

Upvotes: 0

Views: 75

Answers (2)

Vijay Aggarwal
Vijay Aggarwal

Reputation: 148

While my answer is fairly general, I am assuming you are using MySQL.

Short Answer:
Do the following things one by one in the order mentioned while measuring performance improvement with each step.

  1. Add Indexes on tickets.draw and tickets.status. Also add an index (primary key will be even better) on coefficients.number.
  2. Use int instead of varchar wherever possible.
  3. Convert query to stored procedure to save values of FIELD calls and reuse these values in GREATEST instead of calling FIELD again with same values.
  4. Move calls to FIELD at INSERT/UPDATE time instead of SELECT.

Yes, you can update all rows through query. Use your SELECT query and make the following changes to it:

  1. Replace SELECT t.idtickets,t.uplata, c.coefficient FROM by UPDATE
  2. Add SET t.wonamount = c.coefficient*t.uplata, t.status='2' before WHERE ...

(Really) Long Answer:
Your question is a very good case for discussing SQL optimization as there are many optimization techniques that can be applied here. Let me discuss them in increasing order of complexity, so that you can implement them one by one till you are happy with the results. I will also generalize every point for community's benefit while giving precise suggestions to you. Let's start:

  1. All SQL optimization starts with EXPLAIN. It's a sort of black magic that tells what's wrong with your query. Simply add the EXPLAIN keyword before the SELECT keyword in your query and you get a wealth of information on how your query is executed behind the scene. Here is the EXPLAIN output of your query (some fields removed for sake of brevity): 

    +-------+-------+---------------+---------+-------+------+-----------------+    
| table | type  | possible_keys | key     | ref   | rows | Extra           |    
+-------+-------+---------------+---------+-------+------+-----------------+    
| d     | const | PRIMARY       | PRIMARY | const |    1 |                 |    
| t     | ALL   | NULL          | NULL    | NULL  |   13 | Using where     |    
| c     | ALL   | NULL          | NULL    | NULL  |   35 | Using where;... |    
+-------+-------+---------------+---------+-------+------+-----------------+

    Each row covers a table involved in your query. Two important fields to look at here are key and rows. rows tells the number of rows of that table scanned for the query. The more this number, the more data MySQL has to scan, and therefore the slower your query. key tells if MySQL is using any shortcut to reduce rows. In the absence of any key, MySQL has to scan all rows of that table. So, we need to supply keys (also called indexes) to MySQL so that it can reduce rows and execute queries fast.

    • Here, table t (i.e. tickets) is not using any key and therefore scanning all rows (there are 13 rows in the sample data you provided in your fiddle, and 500,000 of them in the real data). So, we add keys (or indexes) to those fields of tickets table that are involved in decision making in this query. These fields are draw and status (... WHERE t.draw='1' AND t.status = '1'...).
      • mysql> ALTER TABLE tickets ADD INDEX idx_draw(draw);
      • mysql> ALTER TABLE tickets ADD INDEX idx_status(status);
    • Similarly, coefficients will benefit by index on number. A PRIMARY KEY on number will be even better if numbers are unique.
  2. Integer data types (short, int, long, etc.) are significantly faster than character data types (char, varchar, etc.). So, avoid using character data types for integer data. In your data, all fields in draws table, and almost all fields in tickets table contain numeric data. (Booleans can be stored as byte instead of varchar. Also consider storing timestamps as int or long instead of varchar.)
  3. FIELD is a costly call, especially if given a lot of arguments, as has to do a lot of work. In your query there are six distinct FIELD calls, and each is repeated in the call to GREATEST function, making 12 calls in total. Consider using stored procedures which allow you to save results of function calls in variables and reuse them later.
  4. Performing validations during INSERT/UPDATE is better than performing them during SELECT. Consider validating your tickets.b1-b6 against draws.1-35 while inserting/updating instead of querying and your SELECT query will be much simpler and faster. The result of GREATEST can also be calculated at insert/update time and saved in an extra field in the tickets table to avoid recalculation every time during SELECT.

As with all queries, your query may need more optimizations when your data grows 100-1000 times its current size, but these should be enough for now.

Upvotes: 2

bart2puck
bart2puck

Reputation: 2522

does your db have indexes? MYSQL indexes

Upvotes: 1

Related Questions