SQL reported an unusual error, which confused the new interns

This article is shared from Huawei cloud community 《

Remember a mysql Format conflict of association query 【 May 04】

》, author ： KevinQ .

Origin of problem

As CRUD The programmer , Of course, the most commonly used function is database query .

A few days ago, a very simple SQL Reported an unusual error ：

Illegal mix of collations (utf8mb4_general_ci,IMPLICIT) and (utf8mb4_unicode_ci,IMPLICIT) for operation '='

Let's have a look today mysql Of utf8mb4 Two format problems , And how to solve this problem ？

mysql Of utf8 Format and sorting rules

Character set uft8 And uft8mb4

stay mysql When creating tables and adding fields in , Want to choose utf8 There are two options for character set ：

• utf8

• utf8mb4

What's the difference between the two ？

MySQL Is in 5.5.3 Then there was utf8mb4 Optional character set ,mb4 It means ：Most Bytes 4, Compatible unicode.

and utf8 Most support 3 Bytes , such as 3 Bytes of uft8 Unable to support Emoji Expressions and unusual Chinese characters , And any new Unicode Character etc. , That's why uft8mb4.

Sort rule

Our problem this time is caused by collation conflict ：

Query statement ：

SELECT * FROM table1 t1 
left join table2 on t1.username = t2.username

The coded character sets of the associated fields are utf8mb4, however t1.username The sorting rule is utf8mb4_general_ci, and t2.username The sorting rule is utf8mb4_unicode_ci.

What are the differences and characteristics between the two sorting rules ？

characteristic

utf8mb4_unicode_ci  It's based on standards  Unicode  To sort and compare , That is, it can support all Unicode Precise sorting of characters ;

and utf8mb4_general_ci It didn't come true Unicode Sort rule , When encountering special characters , The order may be different .

difference

Therefore ,uft8mb4_general_ci comparison utf8mb4_unicode_ci, Although the accuracy of the former is discounted , But comparison and sorting are performed faster , also , Usually, the order in which special characters are encountered is not important .

Solution to collation incompatibility

stay mysql in , There are two solutions to collation incompatibility ：

1、 The first one is , Modify the format of table fields to be consistent .

That is, the sorting rules of the modified table are unified as utf8mb4_unicode_ci, Or another . It is more recommended to use utf8mb4_unicode_ci Collation of , To borrow StackOverflow Last paragraph ：

There is almost certainly no reason to use utf8mb4_general_ci anymore, as we have left behind the point where CPU speed is low enough that the performance difference would be important. Your database will almost certainly be limited by other bottlenecks than this.

Probably means to say , At present CPU The running speed of is fast enough that we can no longer consider this sort as a parameter , And other expenses should be opened .

2、 Another way is to use keywords ,COLLATE

adopt COLLATE attribute , You can specify how columns are sorted and compared .

When we use , Put it where the collation needs to be modified in the association query ：

SELECT
u.guid
FROM
`test` t
LEFT JOIN user u ON u.guid = t.guid COLLATE utf8mb4_unicode_ci
WHERE t.state = 1

such , Use COLLATE attribute , You can use fields with different sorting rules to perform associated queries .

however , After testing , It will slow down like this SQL Query speed .

What is the specific way , Need careful consideration .

Click to follow , The first time to learn about Huawei's new cloud technology ~