json 全文索引 mysql_MySQL全文索引、联合索引、like查询、json查询速度大比拼

我不是女神ヾ 2022-11-07 15:57 250阅读 0赞

查询背景

有一个表tmp\_test\_course大概有10万条记录，然后有个json字段叫outline，存了一对多关系(保存了多个编码，例如jy1577683381775)

我们需要在这10万条数据中检索特定类型的数据，目标总数据量：2931条

SELECT COUNT(\*) FROM tmp\_test\_course WHERE \`type\`=5 AND del=2 AND is\_leaf=1

![ce4d2a7441c729d400b3f576a8b67625.png][]

我们在限定为上面类型的同时，还得包含下面任意一个编码(也就是OR查询)

jy1577683381775

jy1577683380808

jy1577683379178

jy1577683378676

jy1577683377617

jy1577683376672

jy1577683375903

jy1578385720787

jy1499916986208

jy1499917112460

jy1499917093400

jy1499917335579

jy1499917334770

jy1499917333339

jy1499917331557

jy1499917330833

jy1499917329615

jy1499917328496

jy1576922006950

jy1499916993558

jy1499916992308

jy1499917003454

jy1499917002952

下面分别列出4种方式查询outline字段，给出相应的查询时间和扫描行数

一、like查询

耗时248毫秒

SELECT \* FROM tmp\_test\_course

WHERE \`type\`=5 AND del=2 AND is\_leaf=1

AND (

outline like '%jy1577683381775%'

OR outline like '%jy1577683380808%'

OR outline like '%jy1577683379178%'

OR outline like '%jy1577683378676%'

OR outline like '%jy1577683377617%'

OR outline like '%jy1577683376672%'

OR outline like '%jy1577683375903%'

OR outline like '%jy1578385720787%'

OR outline like '%jy1499916986208%'

OR outline like '%jy1499917112460%'

OR outline like '%jy1499917093400%'

OR outline like '%jy1499917335579%'

OR outline like '%jy1499917334770%'

OR outline like '%jy1499917333339%'

OR outline like '%jy1499917331557%'

OR outline like '%jy1499917330833%'

OR outline like '%jy1499917329615%'

OR outline like '%jy1499917328496%'

OR outline like '%jy1576922006950%'

OR outline like '%jy1499916993558%'

OR outline like '%jy1499916992308%'

OR outline like '%jy1499917003454%'

OR outline like '%jy1499917002952%'

)

![70eccdcaebc8eb825841a2f7e4d34382.png][]

EXPLAIN分析结果如下，全表扫描

![3bbded87acdb12e413d1834129c89289.png][]

二、json函数查询

使用函数JSON\_SEARCH，更多函数请查看MySQL官方文档

可以看到，查询耗时196毫秒，速度稍微快了一点

SELECT \* FROM tmp\_test\_course

WHERE \`type\`=5 AND del=2 AND is\_leaf=1

AND

(

JSON\_SEARCH(outline, 'one', 'jy1577683381775') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1577683380808') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1577683379178') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1577683378676') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1577683377617') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1577683376672') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1577683375903') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1578385720787') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499916986208') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499917112460') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499917093400') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499917335579') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499917334770') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499917333339') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499917331557') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499917330833') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499917329615') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499917328496') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1576922006950') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499916993558') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499916992308') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499917003454') IS NOT NULL OR

JSON\_SEARCH(outline, 'one', 'jy1499917002952') IS NOT NULL

)

![ff77f6521120d168f24f986ba1bef318.png][]

EXPLAIN分析结果如下，还是全表扫描

![edc2fdb072bd9688758acbcd50676263.png][]

三、联合索引查询

下面为该表建立一个联合索引(本来想建一个type-del-is\_leaf-outline的索引，但是outline字段太长限制，所以只加type-del-is\_leaf的联合索引

ALTER TABLE tmp\_test\_course ADD KEY \`type-del-is\_leaf\` (\`type\`,\`del\`,\`is\_leaf\`)

加入索引后再执行like和json查询，明显提速。

like执行用了136毫秒，json查询用了82.6毫秒，由此可见针对json类型使用json函数查询比like快

![a8d1f53c625280f268fa4fff7cffc169.png][]

![ba04708bd0803d764ea198da0ff7e30b.png][]

EXPLAIN分析结果如下，两者查询扫描的行数都限定在了2931行

![ee9f668aa3bbbbeb20090a236604c2c3.png][]

![88ce77bb663008bb07e7cb9012e5b92d.png][]

四、全文索引查询

因为全文索引只支持CHAR、VARCHAR和TEXT，我们需要把JSON字段定义改一下

ALTER TABLE tmp\_test\_course MODIFY \`outline\` VARCHAR(1024) NOT NULL DEFAULT '\[\]'

添加全文索引

ALTER TABLE tmp\_test\_course ADD FULLTEXT INDEX outline (outline);

现在再来用全文索引进行检索

SELECT \* FROM tmp\_test\_course

WHERE \`type\`=5 AND del=2 AND is\_leaf=1

AND

MATCH(outline) AGAINST ('jy1577683381775 jy1577683380808 jy1577683379178 jy1577683378676 jy1577683377617 jy1577683376672 jy1577683375903 jy1578385720787 jy1499916986208 jy1499917112460 jy1499917093400 jy1499917335579 jy1499917334770 jy1499917333339 jy1499917331557 jy1499917330833 jy1499917329615 jy1499917328496 jy1576922006950 jy1499916993558 jy1499916992308 jy1499917003454 jy1499917002952')

耗时11.6毫秒,速度提升极其明显，可见全文索引的牛逼。

![e125ad64ee49da9cce3f272b9429a516.png][]

EXPLAIN分析结果如下，显示只扫描了一行

![b8816c4f64301c3c36dcad252c2e5c8d.png][]

结论

以下是4种情况的执行结果

全文索引: 11.6ms

联合索引：82.6ms(json)、136ms(like)

json函数查询：196ms

like查询: 248ms

结论：全文索引 > 联合索引 > json函数查询 > like查询

数据量越大，全文索引速度越明显，就10万的量，查询速度大概比直接查询快了20倍左右，如果是百万或千万级别的表，提升差距会更加大，所以有条件还是老老实实用全文索引吧

[ce4d2a7441c729d400b3f576a8b67625.png]: /images/20221023/a158df1005df4966a8bb4b7357f108ed.png
[70eccdcaebc8eb825841a2f7e4d34382.png]: /images/20221023/574a4181d9a048d5a7a640d5991384bf.png
[3bbded87acdb12e413d1834129c89289.png]: /images/20221023/268a7082e9d44d4780cb6a772fa3806c.png
[ff77f6521120d168f24f986ba1bef318.png]: /images/20221023/922b0814724d4b7983af980a258a5995.png
[edc2fdb072bd9688758acbcd50676263.png]: /images/20221023/758cb5db962d441996eeff68381b91c9.png
[a8d1f53c625280f268fa4fff7cffc169.png]: /images/20221023/d89c94d3b6d44a3c9e2b66b33120315e.png
[ba04708bd0803d764ea198da0ff7e30b.png]: /images/20221023/b680dca964bc433eac0f15314a50d9c6.png
[ee9f668aa3bbbbeb20090a236604c2c3.png]: /images/20221023/85d87c9a4bd644e4859cf4f8b22c591c.png
[88ce77bb663008bb07e7cb9012e5b92d.png]: /images/20221023/6705e35c2dc34afc962422a0e7f0f0c9.png
[e125ad64ee49da9cce3f272b9429a516.png]: /images/20221023/5d2cee5ecdfe4ce799814519ee577f45.png
[b8816c4f64301c3c36dcad252c2e5c8d.png]: /images/20221023/95ba82dcc66f465693d2c5f20484c29c.png