MySQL查询WHERE条件慢的大表

Question

I have a table (ip2country) of 1.5 million record and I'm executing the following query which takes more than 4 seconds My table structure

CREATE TABLE `ip2country` (
  `beginIPNum` bigint(20) DEFAULT NULL,
  `endIPNum` bigint(20) DEFAULT NULL,
  `countryId` varchar(4) DEFAULT NULL,
  `countryName` varchar(50) DEFAULT NULL,
  `state` varchar(50) DEFAULT NULL,
  `city` varchar(50) DEFAULT NULL,
  KEY `index1` (`beginIPNum`,`endIPNum`) USING BTREE
) ENGINE=MyISAM DEFAULT CHARSET=utf8
PARTITION BY RANGE  COLUMNS(beginIPNum,endIPNum)
(PARTITION p0 VALUES LESS THAN (16777216,16777216) ENGINE = MyISAM,
 PARTITION p1 VALUES LESS THAN (251658240,251658240) ENGINE = MyISAM,
 PARTITION p2 VALUES LESS THAN (503316480,503316480) ENGINE = MyISAM,
 PARTITION p3 VALUES LESS THAN (754974720,754974720) ENGINE = MyISAM,
 PARTITION p4 VALUES LESS THAN (1006632960,1006632960) ENGINE = MyISAM,
 PARTITION p5 VALUES LESS THAN (1258291200,1258291200) ENGINE = MyISAM,
 PARTITION p6 VALUES LESS THAN (1509949440,1509949440) ENGINE = MyISAM,
 PARTITION p7 VALUES LESS THAN (1761607680,1761607680) ENGINE = MyISAM,
 PARTITION p8 VALUES LESS THAN (2013265920,2013265920) ENGINE = MyISAM,
 PARTITION p9 VALUES LESS THAN (2264924160,2264924160) ENGINE = MyISAM,
 PARTITION p10 VALUES LESS THAN (2516582400,2516582400) ENGINE = MyISAM,
 PARTITION p11 VALUES LESS THAN (2768240640,2768240640) ENGINE = MyISAM,
 PARTITION p12 VALUES LESS THAN (3019898880,3019898880) ENGINE = MyISAM,
 PARTITION p13 VALUES LESS THAN (3271557120,3271557120) ENGINE = MyISAM,
 PARTITION p14 VALUES LESS THAN (3523215360,3523215360) ENGINE = MyISAM,
 PARTITION p15 VALUES LESS THAN (3774873600,3774873600) ENGINE = MyISAM,
 PARTITION p16 VALUES LESS THAN (4294967295,4294967295) ENGINE = MyISAM)

SELECT beginIPNum,endIPNum,countryId,countryName 
FROM sdportallog.ip2country 
WHERE 2130706433 BETWEEN beginIPNum AND endIPNum

Can anyone help me please?

Answer 1

Try

SELECT beginIPNum,endIPNum,countryId,countryName 
FROM sdportallog.ip2country            
WHERE beginIPNum  <= 2130706433 AND endIPNum >= 2130706433 

Answer 2

The query optimization is to create a couple of indexes. The following script will create two indexes on the table, one on each field.

-- Create index on IPFrom
CREATE INDEX index_beginIPNum
    ON sdportallog.ip2country(beginIPNum)

-- Create index on IPTo
CREATE INDEX index_endIPNum
    ON sdportallog.ip2country(endIPNum)

Answer 3

If you are fine with indexes. You can try the following approach:

1. Add partitions
2. Add another column to your table
3. Change query

Partitions : Try to add partitions (you can test RANGE and HASH). Right strategy you can choose by your own simply re-creating table and re-executing query. For HASH you can start for example with 10 partitions. For RANGE you can:

• inet_aton(1.0.0.0) ... inet_aton(255.0.0.0). No need to create 255 partitions.
• if for any group too many countries exists per group X.0.0.0 you can try to narrow partitions and divide by XY0.0

Tables changes : I suppose it is good practice to add partitioning parameter as column and index it. But feel free this step.

Changes related you query : you should take about adding filtering on your hash/range values. For example, if you are partitioning by X.0.0.0 you should add something like where your_added_column = inet_aton(X.0.0.0) and your_base_condition

Try to play with different partition types, amount of partitions.

Hope it helps.

Answer 4

从sdportallog.ip2country中选择beginIPNum，endIPNum，countryId，countryName WHERE 2130706433 beginIPNum和endIPNum 限制1之间