How MySQL queries and modifies JSON data

Catalog

Yes json Data query and modification

Several related functions

Example

Inquire about

Let's see the modification

Delete

Insert

Mysql Handle json data

Yes json Data query and modification

Use Field ->'$.json attribute ' Query criteria

Use json_extract Function query ,json_extract( Field , "$.json attribute ")

according to json Array query , use JSON_CONTAINS( Field , JSON_OBJECT('json attribute ', " Content ")) : [{}] Query this form json Array

MySQL5.7 Above support JSON The operation of , And added JSON Storage type

General database storage JSON Type of data will use JSON Type or TEXT type

Several related functions

Example

I didn't create it here json Field format of , It USES text Storage json .

Be careful ： use JSON Type words 1）JSON Columns must store JSON Format data , Otherwise, an error will be reported .2）JSON There are no default values for data types .

Insert json Format data into this column ：

{"age": "28", "pwd": "lisi", "name": " Li Si "} Inquire about

1、

select * from `offcn_off_main` where json_extract(json_field,"$.name") = ' Li Si '

2、

select * from `offcn_off_main` where json_field->'$.name' = ' Li Si ' 

Use explain You can see that the index cannot be used .

So it needs to be modified ：

mysql Native does not support json Attribute index in column , But we can pass mysql The virtual column of is indirectly json Create an index for some attributes in , The principle is to json Create virtual columns with attributes in , Then by indexing the virtual columns , Thus, the attribute is indirectly indexed .

stay MySQL 5.7 in , Two kinds of support Generated Column, namely Virtual Generated Column and Stored Generated Column, The former will only Generated Column In the data dictionary （ Table metadata ）, It doesn't persist this column of data to disk ; The latter will Generated Column Persist to disk , It's not calculated every time it's read . Obviously , The latter stores data that can be calculated from existing data , Need more disk space , And Virtual Column There is no advantage in comparison ----（ In fact, I think there are still advantages. After all, there will be less query and calculation ）

therefore ,MySQL 5.7 in , Don't specify Generated Column The type of , The default is Virtual Column.

if necessary Stored Generated Golumn Words , May be in Virtual Generated Column It's more appropriate to index on , In general , All use Virtual Generated Column, This is also MySQL Default method .

The format is as follows ：

fieldname <type> [ GENERATED ALWAYS ] AS ( <expression> ) [ VIRTUAL|STORED ] [ UNIQUE [KEY] ] [ [PRIMARY] KEY ] [ NOT NULL ] [ COMMENT <text> ]

So here I am ：

ALTER TABLE 'off_main' `names_virtual` VARCHAR(20) GENERATED ALWAYS AS (`json_field` ->> '$.name') not null;

Note: Using operators -» To quote JSON In the field KEY. In this case, the field names_virtual For virtual fields , I define it as not null . In real work , We must set specific conditions . because JSON It is a data object with weak structure . In other words, its structure is not fixed .

Index Virtual fields ：

CREATE INDEX `names` ON `off_main`(`names_virtual`);

Note that if the virtual field is not created, the table is added , It's added later , When adding an index, if the virtual field in some rows is null, But it is also set that it cannot be null, Then the index cannot be created successfully , Tips column can not be null.

After adding index explain You can see that the index is used , And the value of the virtual field will follow json The attribute of the field changes automatically .

Let's see the modification update off_main set json_field = json_set(json_field,'$.phone', '132') WHERE id = 45 // Modify multiple at the same time UPDATE offcn_off_main set json_field = json_set(json_field,'$.name',456,'$.age','bbb') WHERE id = 45

json_set() Methods that exist will override , Nonexistent ones will be added .

Delete UPDATE offcn_off_main set json_field = json_remove(json_field,'$.pwd','$.phone') WHERE id = 45 Insert UPDATE offcn_off_main set json_field = json_insert(json_field,'$.pwd','111') WHERE id = 45

insert And update The difference is insert What does not exist will increase , What exists will not cover

Mysql Handle json data

1. If the data is small , take json Data is copied directly to mysql Of json Field , If the data is too large, you can pass java Wait for the backstage form to json Data analysis , Then write to the database .

Query operation  

select *,json->'$.features[0].geometry.rings' as rings from JSON;

Read some data from one table and store it in another table （ A piece of data ）

insert into DT_village(name, border) SELECT  json->'$.features[0].attributes.CJQYMC',json->'$.features[0].geometry.rings'from JSON;

Read json Data and write to the database ( At this time, we use the form of defining function to execute the method , You can define stool volume )

# clear database TRUNCATE table DT_village;# Define stored procedures delimiter //DROP PROCEDURE IF EXISTS insert_test_val;##num_limit Number of data to insert ,rand_limit The largest random number CREATE PROCEDURE insert_test_val() BEGIN DECLARE i int default 0; DECLARE a,b varchar(5000); WHILE i<10 do set a=CONCAT('$.features[',i,'].attributes.CJQYMC'); set b=CONCAT('$.features[',i,'].geometry.rings'); insert into DT_village(name, border) select #json->'$.features[0].attributes.CJQYMC',json->'$.features[0].geometry.rings' # (json->a),(json->b) json_extract(json,a),json_extract(json,b) from JSON; set i = i + 1; END WHILE; END//# Calling stored procedure call insert_test_val();

Call the cursor to get jsosn A row in the data , And perform the insertion operation

delimiter //drop procedure if exists StatisticStore;CREATE PROCEDURE StatisticStore() BEGIN # Create variables to receive cursor data declare j json;# Storage json data DECLARE i int default 0; # Create a total variable , Record the number of executions , Control cycle DECLARE a,b,c varchar(5000);# Definition json The key value of a data in the array # Create an end flag variable declare done int default false; # Create cursors declare cur cursor for select json from JSON where name = '1'; # Specifies the return value at the end of the cursor loop declare continue HANDLER for not found set done = true; # Set initial value set a=CONCAT('$.features[',i,'].attributes.XZQDM'); set b=CONCAT('$.features[',i,'].attributes.XZQMC'); set c=CONCAT('$.features[',i,']'); # Open cursor open cur; # Start looping through the data in the cursor read_loop:loop # According to a piece of data that the cursor currently points to fetch cur into j; # Determine whether the cursor's loop ends if done then leave read_loop;# Jump out of the cursor loop end if; # Here you can do whatever you want WHILE i<11 do insert into dt_border(xzq_code,name,border) select json_extract(j,a),json_extract(j,b),json_extract(j,c) from JSON; set i = i + 1; END WHILE; # End cursor loop end loop; # Close cursor close cur; # Output results select j,i; END;# Calling stored procedure call StatisticStore();

The above is personal experience , I hope I can give you a reference , I also hope you can support the software development network .