我想知道以下内容在性能上是否有区别
SELECT ... FROM ... WHERE someFIELD IN(1,2,3,4)
SELECT ... FROM ... WHERE someFIELD between 0 AND 5
SELECT ... FROM ... WHERE someFIELD = 1 OR someFIELD = 2 OR someFIELD = 3 ...
或者MySQL会像编译器优化代码一样优化SQL吗?
EDIT
根据评论中说明的原因，将AND改为OR。
当前回答
正如其他人解释的那样，就查询性能而言，选择IN比选择OR更好。
在以下情况下，具有OR条件的查询可能需要更长的执行时间。
执行，如果MySQL优化器选择任何其他索引是有效的(在假阳性的情况下)。 如果记录的数量更多(正如Jacob明确表示的那样)
其他回答
下面是使用MySQL 5.6 @SQLFiddle的6个查询的详细信息
总之，这6个查询涵盖了独立索引的列，每个数据类型使用2个查询。所有查询都会使用索引，而不管使用的是in()还是or。
| ORs | IN()
integer | uses index | uses index
date | uses index | uses index
varchar | uses index | uses index
我只是想揭穿“或”表示不能使用索引的说法。这不是真的。索引可以在使用OR的查询中使用，如下例中的6个查询所示。
而且，在我看来，许多人都忽略了一个事实，即IN()是一组or的语法快捷方式。在小范围内，使用IN() -v- OR之间的性能差异是极其微小的。
虽然在更大的范围内，IN()当然更方便，但它在逻辑上仍然等同于一组OR条件。每个查询的环境都会发生变化，因此在表上测试查询总是最好的。
6个解释方案总结，全部“使用索引条件”(右滚动)
Query select_type table type possible_keys key key_len ref rows filtered Extra
------------- --------- ------- --------------- ----------- --------- ----- ------ ---------- -----------------------
Integers using OR SIMPLE mytable range aNum_idx aNum_idx 4 10 100.00 Using index condition
Integers using IN SIMPLE mytable range aNum_idx aNum_idx 4 10 100.00 Using index condition
Dates using OR SIMPLE mytable range aDate_idx aDate_idx 6 7 100.00 Using index condition
Dates using IN SIMPLE mytable range aDate_idx aDate_idx 6 7 100.00 Using index condition
Varchar using OR SIMPLE mytable range aName_idx aName_idx 768 10 100.00 Using index condition
Varchar using IN SIMPLE mytable range aName_idx aName_idx 768 10 100.00 Using index condition
SQL小提琴
MySQL 5.6模式设置:
CREATE TABLE `myTable` (
`id` mediumint(8) unsigned NOT NULL auto_increment,
`aName` varchar(255) default NULL,
`aDate` datetime,
`aNum` mediumint(8),
PRIMARY KEY (`id`)
) AUTO_INCREMENT=1;
ALTER TABLE `myTable` ADD INDEX `aName_idx` (`aName`);
ALTER TABLE `myTable` ADD INDEX `aDate_idx` (`aDate`);
ALTER TABLE `myTable` ADD INDEX `aNum_idx` (`aNum`);
INSERT INTO `myTable` (`aName`,`aDate`)
VALUES
("Daniel","2017-09-19 01:22:31")
,("Quentin","2017-06-03 01:06:45")
,("Chester","2017-06-14 17:49:36")
,("Lev","2017-08-30 06:27:59")
,("Garrett","2018-10-04 02:40:37")
,("Lane","2017-01-22 17:11:21")
,("Chaim","2017-09-20 11:13:46")
,("Kieran","2018-03-10 18:37:26")
,("Cedric","2017-05-20 16:25:10")
,("Conan","2018-07-10 06:29:39")
,("Rudyard","2017-07-14 00:04:00")
,("Chadwick","2018-08-18 08:54:08")
,("Darius","2018-10-02 06:55:56")
,("Joseph","2017-06-19 13:20:33")
,("Wayne","2017-04-02 23:20:25")
,("Hall","2017-10-13 00:17:24")
,("Craig","2016-12-04 08:15:22")
,("Keane","2018-03-12 04:21:46")
,("Russell","2017-07-14 17:21:58")
,("Seth","2018-07-25 05:51:30")
,("Cole","2018-06-09 15:32:53")
,("Donovan","2017-08-12 05:21:35")
,("Damon","2017-06-27 03:44:19")
,("Brian","2017-02-01 23:35:20")
,("Harper","2017-08-25 04:29:27")
,("Chandler","2017-09-30 23:54:06")
,("Edward","2018-07-30 12:18:07")
,("Curran","2018-05-23 09:31:53")
,("Uriel","2017-05-08 03:31:43")
,("Honorato","2018-04-07 14:57:53")
,("Griffin","2017-01-07 23:35:31")
,("Hasad","2017-05-15 05:32:41")
,("Burke","2017-07-04 01:11:19")
,("Hyatt","2017-03-14 17:12:28")
,("Brenden","2017-10-17 05:16:14")
,("Ryan","2018-10-10 08:07:55")
,("Giacomo","2018-10-06 14:21:21")
,("James","2018-02-06 02:45:59")
,("Colt","2017-10-10 08:11:26")
,("Kermit","2017-09-18 16:57:16")
,("Drake","2018-05-20 22:08:36")
,("Berk","2017-04-16 17:39:32")
,("Alan","2018-09-01 05:33:05")
,("Deacon","2017-04-20 07:03:05")
,("Omar","2018-03-02 15:04:32")
,("Thaddeus","2017-09-19 04:07:54")
,("Troy","2016-12-13 04:24:08")
,("Rogan","2017-11-02 00:03:25")
,("Grant","2017-08-21 01:45:16")
,("Walker","2016-11-26 15:54:52")
,("Clarke","2017-07-20 02:26:56")
,("Clayton","2018-08-16 05:09:29")
,("Denton","2018-08-11 05:26:05")
,("Nicholas","2018-07-19 09:29:55")
,("Hashim","2018-08-10 20:38:06")
,("Todd","2016-10-25 01:01:36")
,("Xenos","2017-05-11 22:50:35")
,("Bert","2017-06-17 18:08:21")
,("Oleg","2018-01-03 13:10:32")
,("Hall","2018-06-04 01:53:45")
,("Evan","2017-01-16 01:04:25")
,("Mohammad","2016-11-18 05:42:52")
,("Armand","2016-12-18 06:57:57")
,("Kaseem","2018-06-12 23:09:57")
,("Colin","2017-06-29 05:25:52")
,("Arthur","2016-12-29 04:38:13")
,("Xander","2016-11-14 19:35:32")
,("Dante","2016-12-01 09:01:04")
,("Zahir","2018-02-17 14:44:53")
,("Raymond","2017-03-09 05:33:06")
,("Giacomo","2017-04-17 06:12:52")
,("Fulton","2017-06-04 00:41:57")
,("Chase","2018-01-14 03:03:57")
,("William","2017-05-08 09:44:59")
,("Fuller","2017-03-31 20:35:20")
,("Jarrod","2017-02-15 02:45:29")
,("Nissim","2018-03-11 14:19:25")
,("Chester","2017-11-05 00:14:27")
,("Perry","2017-12-24 11:58:04")
,("Theodore","2017-06-26 12:34:12")
,("Mason","2017-10-02 03:53:49")
,("Brenden","2018-10-08 10:09:47")
,("Jerome","2017-11-05 20:34:25")
,("Keaton","2018-08-18 00:55:56")
,("Tiger","2017-05-21 16:59:07")
,("Benjamin","2018-04-10 14:46:36")
,("John","2018-09-05 18:53:03")
,("Jakeem","2018-10-11 00:17:38")
,("Kenyon","2017-12-18 22:19:29")
,("Ferris","2017-03-29 06:59:13")
,("Hoyt","2017-01-03 03:48:56")
,("Fitzgerald","2017-07-27 11:27:52")
,("Forrest","2017-10-05 23:14:21")
,("Jordan","2017-01-11 03:48:09")
,("Lev","2017-05-25 08:03:39")
,("Chase","2017-06-18 19:09:23")
,("Ryder","2016-12-13 12:50:50")
,("Malik","2017-11-19 15:15:55")
,("Zeph","2018-04-04 11:22:12")
,("Amala","2017-01-29 07:52:17")
;
.
update MyTable
set aNum = id
;
查询1:
select 'aNum by OR' q, mytable.*
from mytable
where aNum = 12
OR aNum = 22
OR aNum = 27
OR aNum = 32
OR aNum = 42
OR aNum = 52
OR aNum = 62
OR aNum = 65
OR aNum = 72
OR aNum = 82
结果:
| q | id | aName | aDate | aNum |
|------------|----|----------|----------------------|------|
| aNum by OR | 12 | Chadwick | 2018-08-18T08:54:08Z | 12 |
| aNum by OR | 22 | Donovan | 2017-08-12T05:21:35Z | 22 |
| aNum by OR | 27 | Edward | 2018-07-30T12:18:07Z | 27 |
| aNum by OR | 32 | Hasad | 2017-05-15T05:32:41Z | 32 |
| aNum by OR | 42 | Berk | 2017-04-16T17:39:32Z | 42 |
| aNum by OR | 52 | Clayton | 2018-08-16T05:09:29Z | 52 |
| aNum by OR | 62 | Mohammad | 2016-11-18T05:42:52Z | 62 |
| aNum by OR | 65 | Colin | 2017-06-29T05:25:52Z | 65 |
| aNum by OR | 72 | Fulton | 2017-06-04T00:41:57Z | 72 |
| aNum by OR | 82 | Brenden | 2018-10-08T10:09:47Z | 82 |
查询2:
select 'aNum by IN' q, mytable.*
from mytable
where aNum IN (
12
, 22
, 27
, 32
, 42
, 52
, 62
, 65
, 72
, 82
)
结果:
| q | id | aName | aDate | aNum |
|------------|----|----------|----------------------|------|
| aNum by IN | 12 | Chadwick | 2018-08-18T08:54:08Z | 12 |
| aNum by IN | 22 | Donovan | 2017-08-12T05:21:35Z | 22 |
| aNum by IN | 27 | Edward | 2018-07-30T12:18:07Z | 27 |
| aNum by IN | 32 | Hasad | 2017-05-15T05:32:41Z | 32 |
| aNum by IN | 42 | Berk | 2017-04-16T17:39:32Z | 42 |
| aNum by IN | 52 | Clayton | 2018-08-16T05:09:29Z | 52 |
| aNum by IN | 62 | Mohammad | 2016-11-18T05:42:52Z | 62 |
| aNum by IN | 65 | Colin | 2017-06-29T05:25:52Z | 65 |
| aNum by IN | 72 | Fulton | 2017-06-04T00:41:57Z | 72 |
| aNum by IN | 82 | Brenden | 2018-10-08T10:09:47Z | 82 |
查询3:
select 'adate by OR' q, mytable.*
from mytable
where aDate= str_to_date("2017-02-15 02:45:29",'%Y-%m-%d %h:%i:%s')
OR aDate = str_to_date("2018-03-10 18:37:26",'%Y-%m-%d %h:%i:%s')
OR aDate = str_to_date("2017-05-20 16:25:10",'%Y-%m-%d %h:%i:%s')
OR aDate = str_to_date("2018-07-10 06:29:39",'%Y-%m-%d %h:%i:%s')
OR aDate = str_to_date("2017-07-14 00:04:00",'%Y-%m-%d %h:%i:%s')
OR aDate = str_to_date("2018-08-18 08:54:08",'%Y-%m-%d %h:%i:%s')
OR aDate = str_to_date("2018-10-02 06:55:56",'%Y-%m-%d %h:%i:%s')
OR aDate = str_to_date("2017-04-20 07:03:05",'%Y-%m-%d %h:%i:%s')
OR aDate = str_to_date("2018-03-02 15:04:32",'%Y-%m-%d %h:%i:%s')
OR aDate = str_to_date("2017-09-19 04:07:54",'%Y-%m-%d %h:%i:%s')
OR aDate = str_to_date("2016-12-13 04:24:08",'%Y-%m-%d %h:%i:%s')
结果:
| q | id | aName | aDate | aNum |
|-------------|----|----------|----------------------|------|
| adate by OR | 47 | Troy | 2016-12-13T04:24:08Z | 47 |
| adate by OR | 76 | Jarrod | 2017-02-15T02:45:29Z | 76 |
| adate by OR | 44 | Deacon | 2017-04-20T07:03:05Z | 44 |
| adate by OR | 46 | Thaddeus | 2017-09-19T04:07:54Z | 46 |
| adate by OR | 10 | Conan | 2018-07-10T06:29:39Z | 10 |
| adate by OR | 12 | Chadwick | 2018-08-18T08:54:08Z | 12 |
| adate by OR | 13 | Darius | 2018-10-02T06:55:56Z | 13 |
查询4:
select 'adate by IN' q, mytable.*
from mytable
where aDate IN (
str_to_date("2017-02-15 02:45:29",'%Y-%m-%d %h:%i:%s')
, str_to_date("2018-03-10 18:37:26",'%Y-%m-%d %h:%i:%s')
, str_to_date("2017-05-20 16:25:10",'%Y-%m-%d %h:%i:%s')
, str_to_date("2018-07-10 06:29:39",'%Y-%m-%d %h:%i:%s')
, str_to_date("2017-07-14 00:04:00",'%Y-%m-%d %h:%i:%s')
, str_to_date("2018-08-18 08:54:08",'%Y-%m-%d %h:%i:%s')
, str_to_date("2018-10-02 06:55:56",'%Y-%m-%d %h:%i:%s')
, str_to_date("2017-04-20 07:03:05",'%Y-%m-%d %h:%i:%s')
, str_to_date("2018-03-02 15:04:32",'%Y-%m-%d %h:%i:%s')
, str_to_date("2017-09-19 04:07:54",'%Y-%m-%d %h:%i:%s')
, str_to_date("2016-12-13 04:24:08",'%Y-%m-%d %h:%i:%s')
)
结果:
| q | id | aName | aDate | aNum |
|-------------|----|----------|----------------------|------|
| adate by IN | 47 | Troy | 2016-12-13T04:24:08Z | 47 |
| adate by IN | 76 | Jarrod | 2017-02-15T02:45:29Z | 76 |
| adate by IN | 44 | Deacon | 2017-04-20T07:03:05Z | 44 |
| adate by IN | 46 | Thaddeus | 2017-09-19T04:07:54Z | 46 |
| adate by IN | 10 | Conan | 2018-07-10T06:29:39Z | 10 |
| adate by IN | 12 | Chadwick | 2018-08-18T08:54:08Z | 12 |
| adate by IN | 13 | Darius | 2018-10-02T06:55:56Z | 13 |
查询5:
select 'name by OR' q, mytable.*
from mytable
where aname = 'Alan'
OR aname = 'Brian'
OR aname = 'Chandler'
OR aname = 'Darius'
OR aname = 'Evan'
OR aname = 'Ferris'
OR aname = 'Giacomo'
OR aname = 'Hall'
OR aname = 'James'
OR aname = 'Jarrod'
结果:
| q | id | aName | aDate | aNum |
|-------------|----|----------|----------------------|------|
| name by OR | 43 | Alan | 2018-09-01T05:33:05Z | 43 |
| name by OR | 24 | Brian | 2017-02-01T23:35:20Z | 24 |
| name by OR | 26 | Chandler | 2017-09-30T23:54:06Z | 26 |
| name by OR | 13 | Darius | 2018-10-02T06:55:56Z | 13 |
| name by OR | 61 | Evan | 2017-01-16T01:04:25Z | 61 |
| name by OR | 90 | Ferris | 2017-03-29T06:59:13Z | 90 |
| name by OR | 37 | Giacomo | 2018-10-06T14:21:21Z | 37 |
| name by OR | 71 | Giacomo | 2017-04-17T06:12:52Z | 71 |
| name by OR | 16 | Hall | 2017-10-13T00:17:24Z | 16 |
| name by OR | 60 | Hall | 2018-06-04T01:53:45Z | 60 |
| name by OR | 38 | James | 2018-02-06T02:45:59Z | 38 |
| name by OR | 76 | Jarrod | 2017-02-15T02:45:29Z | 76 |
查询6:
select 'name by IN' q, mytable.*
from mytable
where aname IN (
'Alan'
,'Brian'
,'Chandler'
, 'Darius'
, 'Evan'
, 'Ferris'
, 'Giacomo'
, 'Hall'
, 'James'
, 'Jarrod'
)
结果:
| q | id | aName | aDate | aNum |
|------------|----|----------|----------------------|------|
| name by IN | 43 | Alan | 2018-09-01T05:33:05Z | 43 |
| name by IN | 24 | Brian | 2017-02-01T23:35:20Z | 24 |
| name by IN | 26 | Chandler | 2017-09-30T23:54:06Z | 26 |
| name by IN | 13 | Darius | 2018-10-02T06:55:56Z | 13 |
| name by IN | 61 | Evan | 2017-01-16T01:04:25Z | 61 |
| name by IN | 90 | Ferris | 2017-03-29T06:59:13Z | 90 |
| name by IN | 37 | Giacomo | 2018-10-06T14:21:21Z | 37 |
| name by IN | 71 | Giacomo | 2017-04-17T06:12:52Z | 71 |
| name by IN | 16 | Hall | 2017-10-13T00:17:24Z | 16 |
| name by IN | 60 | Hall | 2018-06-04T01:53:45Z | 60 |
| name by IN | 38 | James | 2018-02-06T02:45:59Z | 38 |
| name by IN | 76 | Jarrod | 2017-02-15T02:45:29Z | 76 |
就在你以为安全的时候…
你的eq_range_index_dive_limit值是多少?特别地，你在In子句中有更多还是更少的条目?
这将不包括基准测试，但将深入了解内部工作原理。让我们使用一个工具来查看发生了什么——Optimizer Trace。
查询:SELECT * FROM canada WHERE id…
使用3个OR值，部分跟踪看起来像这样:
"condition_processing": {
"condition": "WHERE",
"original_condition": "((`canada`.`id` = 296172) or (`canada`.`id` = 295093) or (`canada`.`id` = 293626))",
"steps": [
{
"transformation": "equality_propagation",
"resulting_condition": "(multiple equal(296172, `canada`.`id`) or multiple equal(295093, `canada`.`id`) or multiple equal(293626, `canada`.`id`))"
},
...
"analyzing_range_alternatives": {
"range_scan_alternatives": [
{
"index": "id",
"ranges": [
"293626 <= id <= 293626",
"295093 <= id <= 295093",
"296172 <= id <= 296172"
],
"index_dives_for_eq_ranges": true,
"chosen": true
...
"refine_plan": [
{
"table": "`canada`",
"pushed_index_condition": "((`canada`.`id` = 296172) or (`canada`.`id` = 295093) or (`canada`.`id` = 293626))",
"table_condition_attached": null,
"access_type": "range"
}
]
注意颅内压是如何给予ORs的。这意味着OR没有转换成IN, InnoDB将通过ICP执行一堆=测试。(我觉得不值得考虑MyISAM。)
(这是Percona的5.6.22-71.0-log;Id是二级索引。)
现在来看带有几个值的IN()
Eq_range_index_dive_limit = 10;有8个值。
"condition_processing": {
"condition": "WHERE",
"original_condition": "(`canada`.`id` in (296172,295093,293626,295573,297148,296127,295588,295810))",
"steps": [
{
"transformation": "equality_propagation",
"resulting_condition": "(`canada`.`id` in (296172,295093,293626,295573,297148,296127,295588,295810))"
},
...
"analyzing_range_alternatives": {
"range_scan_alternatives": [
{
"index": "id",
"ranges": [
"293626 <= id <= 293626",
"295093 <= id <= 295093",
"295573 <= id <= 295573",
"295588 <= id <= 295588",
"295810 <= id <= 295810",
"296127 <= id <= 296127",
"296172 <= id <= 296172",
"297148 <= id <= 297148"
],
"index_dives_for_eq_ranges": true,
"chosen": true
...
"refine_plan": [
{
"table": "`canada`",
"pushed_index_condition": "(`canada`.`id` in (296172,295093,293626,295573,297148,296127,295588,295810))",
"table_condition_attached": null,
"access_type": "range"
}
]
注意，IN似乎没有被转换成OR。
附注:注意，常数值是排序的。这在两个方面是有益的:
通过更少的跳跃，可能会有更好的缓存，更少的I/O来获取所有的值。 如果两个类似的查询来自不同的连接，并且它们位于事务中，则更有可能导致延迟，而不是由于重叠列表而导致死锁。
最后，有很多值的IN()
{
"condition_processing": {
"condition": "WHERE",
"original_condition": "(`canada`.`id` in (293831,292259,292881,293440,292558,295792,292293,292593,294337,295430,295034,297060,293811,295587,294651,295559,293213,295742,292605,296018,294529,296711,293919,294732,294689,295540,293000,296916,294433,297112,293815,292522,296816,293320,293232,295369,291894,293700,291839,293049,292738,294895,294473,294023,294173,293019,291976,294923,294797,296958,294075,293450,296952,297185,295351,295736,296312,294330,292717,294638,294713,297176,295896,295137,296573,292236,294966,296642,296073,295903,293057,294628,292639,293803,294470,295353,297196,291752,296118,296964,296185,295338,295956,296064,295039,297201,297136,295206,295986,292172,294803,294480,294706,296975,296604,294493,293181,292526,293354,292374,292344,293744,294165,295082,296203,291918,295211,294289,294877,293120,295387))",
"steps": [
{
"transformation": "equality_propagation",
"resulting_condition": "(`canada`.`id` in (293831,292259,292881,293440,292558,295792,292293,292593,294337,295430,295034,297060,293811,295587,294651,295559,293213,295742,292605,296018,294529,296711,293919,294732,294689,295540,293000,296916,294433,297112,293815,292522,296816,293320,293232,295369,291894,293700,291839,293049,292738,294895,294473,294023,294173,293019,291976,294923,294797,296958,294075,293450,296952,297185,295351,295736,296312,294330,292717,294638,294713,297176,295896,295137,296573,292236,294966,296642,296073,295903,293057,294628,292639,293803,294470,295353,297196,291752,296118,296964,296185,295338,295956,296064,295039,297201,297136,295206,295986,292172,294803,294480,294706,296975,296604,294493,293181,292526,293354,292374,292344,293744,294165,295082,296203,291918,295211,294289,294877,293120,295387))"
},
...
"analyzing_range_alternatives": {
"range_scan_alternatives": [
{
"index": "id",
"ranges": [
"291752 <= id <= 291752",
"291839 <= id <= 291839",
...
"297196 <= id <= 297196",
"297201 <= id <= 297201"
],
"index_dives_for_eq_ranges": false,
"rows": 111,
"chosen": true
...
"refine_plan": [
{
"table": "`canada`",
"pushed_index_condition": "(`canada`.`id` in (293831,292259,292881,293440,292558,295792,292293,292593,294337,295430,295034,297060,293811,295587,294651,295559,293213,295742,292605,296018,294529,296711,293919,294732,294689,295540,293000,296916,294433,297112,293815,292522,296816,293320,293232,295369,291894,293700,291839,293049,292738,294895,294473,294023,294173,293019,291976,294923,294797,296958,294075,293450,296952,297185,295351,295736,296312,294330,292717,294638,294713,297176,295896,295137,296573,292236,294966,296642,296073,295903,293057,294628,292639,293803,294470,295353,297196,291752,296118,296964,296185,295338,295956,296064,295039,297201,297136,295206,295986,292172,294803,294480,294706,296975,296604,294493,293181,292526,293354,292374,292344,293744,294165,295082,296203,291918,295211,294289,294877,293120,295387))",
"table_condition_attached": null,
"access_type": "range"
}
]
旁注:我需要这个，由于跟踪的体积:
@@global.optimizer_trace_max_mem_size = 32222;
我敢打赌它们是一样的，你可以通过执行以下操作来运行测试:
循环遍历“in(1,2,3,4)”500次，看看需要多长时间。循环“=1 or= 2 or=3…”版本500次，看看它运行了多长时间。
你也可以尝试一个连接的方式，如果someField是一个索引，你的表很大，它可以更快…
SELECT ...
FROM ...
INNER JOIN (SELECT 1 as newField UNION ALL SELECT 2 UNION ALL SELECT 3 UNION ALL SELECT 4) dt ON someFIELD =newField
我在我的SQL Server上尝试了上面的join方法，它几乎与in(1,2,3,4)相同，它们都导致了一个聚集索引查找。我不确定MySQL将如何处理它们。
我认为对sunseeker的观察的一个解释是，MySQL实际上对in语句中的值进行排序，如果它们都是静态值，并且使用二进制搜索，这比普通或替代方法更有效。我不记得在哪里读到过，但圣汐者的结果似乎是一个证明。
2018: IN(…)更快。但是>= && <=比IN还要快。
这是我的基准。
