如何创建触发器以将更改事件添加到审核日志表中

Posted 2023-03-26

技术标签:

【中文标题】如何创建触发器以将更改事件添加到审核日志表中【英文标题】：How to create Triggers to add the change events into Audit Log tables 【发布时间】：2013-02-27 17:52:51 【问题描述】：

假设我们在数据库中有50 tables，并且我们想要捕获每个表的列中的所有更改（列的先前值和新值）。将有一个审核表，其中包含以下列：

ID、Server_Name、User_Name、Date_Time、Table_Name、Column_Name、Old_Value、New_Value

将有一个审计表，它将捕获该数据库中所有表的更改。我相信我们可以为该数据库的每个表创建触发器。但是请让我知道如何将所有数据添加到一个审计表中。如果您能给我提供一个非常有用的工作示例。

感谢和问候，帕萨

【问题讨论】：

【参考方案1】：

我可以为你提供一种算法，大部分基础工作已经完成：

这可以是您的审计表，应根据您的要求添加时间戳列作为修改日期或更多信息：

CREATE TABLE audit (
     old_data VARCHAR(100),
     new_data VARCHAR(100),
     tbl_name VARCHAR(100)
)
|

这可以用作参考触发器；请注意，每个表都会有一个单独的触发器：

CREATE TRIGGER testtrigger BEFORE UPDATE ON <table_name>
  FOR EACH ROW BEGIN
    INSERT INTO audit(old_data, new_data, tbl_name) VALUES (OLD.first_name, NEW.first_name, "testtable");
  END;
|

每列可以有多个插入语句。如果您想限制不插入未更改的数据，您可以在触发器中进行以下更改：

IF(OLD.column_name <> NEW.column_name) THEN
    --Your insert query here
ELSE
    --NOOP
END IF;

如果需要更多信息，请告知。

【讨论】：

ya .. 我需要有关此的更多信息... 我还想存储列名以及如何动态获取表格名称 @AnishRai，当您为表定义触发器时，您不会随身携带列详细信息。当您插入审计时使用列名。但是 .. 假设我在审计表中存储操作意味着有一个列操作并插入更新 first_name 如果用户在这种情况下更新名字，我怎么知道哪一列被更新用户当您检查数据时，您的操作方式发生了变化，OLD/NEW。，您可以使用相同的 column_name 作为常量。你肯定有不同的数据类型。你不能用 VARCHAR 来存储它们【参考方案2】：

您可以使用此触发器，但如果它适用于每个表对我来说是最好的，因为您可以控制表结构是否发生变化并且不影响其他表，您可以使用此 repo 的示例： https://github.com/areliszxz/mysql_audit

DELIMITER $$
USE `tudbaauditar`$$
CREATE
TRIGGER `tudbaauditar`.`update`
BEFORE UPDATE ON `tudbaauditar`.`tutablaaauditar` #aqui puedes poner antes o despues del update
FOR EACH ROW
BEGIN
        /*Paso de variables para un mejor control*/
        set @res1 = ''; set @res2 = ''; set @res3 = ''; set @res4 = '';
        /*Sacamos info de la ip donde se ejecuta la accion de UPDATE*/
        select host as IP INTO @ipcl from information_schema.processlist WHERE ID=connection_id();
        #concatenamos los campos de la tabla a auditar y verificamos que no sean null, en caso de que los campos sean null agregamos un espacio
        #las variables (new,old)son de mysql, el valor old es el que ya se tenia en la tabla y el new es el valor que se modifico

        #Valores viejos
        SET @oldq = CONCAT (' id ',ifnull(OLD.id,''),
                                                        ' campo1 ',ifnull(OLD.campo1,''),
                                                        ' campo2 ',ifnull(OLD.campo2,''),
                                                        ' campo3 ',ifnull(OLD.campo3,''));
        #Valores nuevos
        SET @newq = CONCAT (' id ',ifnull(new.id,''),
                                                        ' campo1 ',ifnull(new.campo1,''),
                                                        ' campo2 ',ifnull(new.campo2,''),
                                                        ' campo3 ',ifnull(new.campo3,''));
    #guardamos en una variable los valores que unicamente cambiaron                                                 
    IF OLD.id <> new.id THEN set @res1 = CONCAT ('Cambio id ',ifnull(OLD.id,''), ' a: ',ifnull(new.id,'')); END IF;
    IF OLD.campo1 <> new.campo1 THEN set @res2 = CONCAT ('Cambio campo1 ',ifnull(OLD.campo1,''), ' a: ',ifnull(new.campo1,'')); END IF;
    IF OLD.campo2 <> new.campo2 THEN set @res3 = CONCAT ('Cambio campo2 ',ifnull(OLD.campo2,''), ' a: ',ifnull(new.campo2,'')); END IF;
    IF OLD.campo3 <> new.campo3 THEN set @res4 = CONCAT ('Cambio campo3 ',ifnull(OLD.campo3,''), ' a: ',ifnull(new.campo3,'')); END IF;
    SET @resC=CONCAT(ifnull(@res1,''),'|',ifnull(@res2,''),'|',ifnull(@res3,''),'|',ifnull(@res4,''));

    #insertamos en nuestra tabla de log la informacion
    INSERT INTO basedeauditoria.tablalogs (old,new,usuario,typo,fecha,tabla,valor_alterado,ip)                
    VALUES (@oldq ,@newq,CURRENT_USER,"UPDATE",NOW(),"tutablaaauditar",ifnull(@resC,'No cambio nada'),@ipcl);
END$$

#log de insertados(Nuevos registros)
DELIMITER $$
USE `tudbaauditar`$$
CREATE
TRIGGER `tudbaauditar`.`incert`
BEFORE INSERT ON `tudbaauditar`.`tutablaaauditar`
FOR EACH ROW
BEGIN
    SET @oldq = '';
    SET @newq = CONCAT (' id ',ifnull(new.id,''),
    ' campo1 ',ifnull(new.campo1,''),
    ' campo2 ',ifnull(new.campo2,''),
    ' campo3 ',ifnull(new.campo3,''));
    INSERT INTO sys_logdev.logs (old,new,usuario,typo,fecha,tabla)                
    VALUES (@oldq ,@newq,CURRENT_USER,"INSERT",NOW(),"tutablaaauditar");
END$$

#log de Borrados
DELIMITER $$
USE `tudbaauditar`$$
CREATE
TRIGGER `tudbaauditar`.`delete`
AFTER DELETE ON `tudbaauditar`.`tutablaaauditar`
FOR EACH ROW
BEGIN
    SET @newq = '';
    SET @oldq = CONCAT (' id ',ifnull(new.id,''),
    ' campo1 ',ifnull(new.campo1,''),
    ' campo2 ',ifnull(new.campo2,''),
    ' campo3 ',ifnull(new.campo3,''));
    INSERT INTO sys_logdev.logs (old,new,usuario,typo,fecha,tabla)                
    VALUES (@oldq ,@newq,CURRENT_USER,"DELETE",NOW(),"tutablaaauditar");
END$$

【讨论】：

Repo README.md 和英文源代码注释会很不错;)【参考方案3】：

这是 Vlad 答案的一点改进版本。审核表包含更改“差异”列。

审核规则：

INSERT 和 DELETE - 包含所有字段的完整记录存储到审计表中 UPDATE - 仅存储字段的更改

审核表结构。 diff 是存储更改的列。

注意：privileges 是我们在这种情况下审核的表上的 JSON 列。

CREATE TABLE roles_audit_log (
    roles_id INT UNSIGNED NOT NULL,
    diff JSON,
    dml_type ENUM('INSERT', 'UPDATE', 'DELETE') NOT NULL,
    dml_timestamp TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
    dml_created_by VARCHAR(255) NOT NULL default 'system',
    PRIMARY KEY (roles_id, dml_type, dml_timestamp)
);

UPDATE 语句的触发器。将使用已更改的字段填充 diff 列。

DELIMITER $$

CREATE TRIGGER roles_audit_au AFTER UPDATE ON `roles` FOR EACH ROW
BEGIN
    DECLARE m_change text;
    SET m_change = JSON_OBJECT();

    IF coalesce(NEW.role, '') != coalesce(OLD.role, '') THEN
        SET m_change = JSON_SET(m_change, '$.role', NEW.role);
    END IF;

    IF coalesce(NEW.created_at, '') != coalesce(OLD.created_at, '') THEN
        SET m_change = JSON_SET(m_change, '$.created_at', NEW.created_at);
    END IF;

    IF coalesce(NEW.updated_at, '') != coalesce(OLD.updated_at, '') THEN
        SET m_change = JSON_SET(m_change, '$.updated_at', NEW.updated_at);
    END IF;

    IF coalesce(NEW.privileges, '') != coalesce(OLD.privileges, '') THEN
        SET m_change = JSON_SET(m_change, '$.privileges', NEW.privileges);
    END IF;

    INSERT INTO `roles_audit_log` (
        roles_id,
        diff,
        dml_type,
        dml_created_by
    ) VALUES(
        NEW.id,
        m_change,
        'UPDATE',
        coalesce(@logged_user, 'system')
    );
END;$$
DELIMITER ;

DELETE 语句的触发器。将使用已删除的完整行填充 diff。

DELIMITER $$

CREATE TRIGGER roles_audit_ad AFTER DELETE ON `roles` FOR EACH ROW
BEGIN
    INSERT INTO `roles_audit_log` (
        roles_id,
        diff,
        dml_type,
        dml_created_by
    ) VALUES(
        OLD.id,
        JSON_OBJECT(
            'id', OLD.id,
            'role', OLD.role,
            'privileges', OLD.privileges,
            'created_at', OLD.created_at,
            'updated_at', OLD.updated_at
        ),
        'DELETE',
        coalesce(@logged_user, 'system')
    );
END;$$
DELIMITER ;

INSERT 语句的触发器。将使用已删除的完整行填充 diff。

DELIMITER $$

CREATE TRIGGER roles_audit_ai AFTER INSERT ON `roles` FOR EACH ROW
BEGIN
    INSERT INTO `roles_audit_log` (
        roles_id,
        diff,
        dml_type,
        dml_created_by
    ) VALUES(
        NEW.id,
        JSON_OBJECT(
            'id', NEW.id,
            'role', NEW.role,
            'privileges', NEW.privileges,
            'created_at', NEW.created_at,
            'updated_at', NEW.updated_at
        ),
        'INSERT',
        coalesce(@logged_user, 'system')
    );
END;$$
DELIMITER ;

【讨论】：

【参考方案4】：

我花了几天时间想出一个存储过程来自动/动态地在 MariaDB（与 v 10.1.9 一起使用）中创建 UPDATE / DELETE 触发器，以审核更新和删除的所有更改。该解决方案使用 INFORMATION_SCHEMA 为您的每个表自动构建审计触发器。在更新时仅审核更改的列，而在删除时，所有历史记录都保留在审核中。

在下面的示例中，我们创建了一个包含两个表的测试数据库，tb_company 和 tb_auditdetail 将保存我们的审计日志。

    -- Dynamic Automated Update / Delete Triggers in MariaDB
    -- Leonard Tonna 19/05/2016 - www.ilabmalta.com

    CREATE DATABASE db_ilabmalta_test;

    USE db_ilabmalta_test;

    CREATE TABLE tb_auditDetail(
        audit_pk int(9) NOT NULL PRIMARY KEY AUTO_INCREMENT,
        type varchar(1) NOT NULL,
        tablename varchar(128) NULL,
        pk varchar(128) NULL,
        fieldname varchar(128) NULL,
        oldvalue varchar(1000) NULL,
        newvalue varchar(1000) NULL,
        updatedate datetime NULL,
        username varchar(128) NULL,
        dbusername varchar(128) NULL,
        machinename varchar(128) NULL);

    CREATE TABLE tb_company(
        cmp_pk int(9) NOT NULL PRIMARY KEY AUTO_INCREMENT,
        cmp_name varchar(100) NOT NULL,
        cmp_no varchar(16) NULL,
        cmp_status smallint NOT NULL DEFAULT 1,
        cmp_created datetime NOT NULL DEFAULT CURRENT_TIMESTAMP,
        cmp_createdby varchar(10) NOT NULL,
        cmp_updated datetime NULL,
        cmp_updatedby varchar(10) NULL,
        cmp_record_version int(9) NOT NULL DEFAULT 1 ) ;

    -- We now create sp_maketrigger which is the stored procedure
    -- which will give us our trigger scripts

    DELIMITER $$

    DROP PROCEDURE IF EXISTS sp_maketrigger; 

    CREATE PROCEDURE sp_maketrigger (IN s_tablename CHAR(30), OUT u_trigger_out VARCHAR(65500) CHARACTER SET ascii,OUT d_trigger_out VARCHAR(65500) CHARACTER SET ascii)
    BEGIN
        DECLARE s_fieldname VARCHAR(50);
        DECLARE u_trigger VARCHAR(65500) CHARACTER SET ascii;
        DECLARE d_trigger VARCHAR(65500) CHARACTER SET ascii;
        DECLARE s_key VARCHAR(50);
        DECLARE s_updatedby VARCHAR(50);
        DECLARE s_updated VARCHAR(50);
        DECLARE s_recversion VARCHAR(50);
        DECLARE done INT DEFAULT 0; 
        DECLARE cursor_end CONDITION FOR SQLSTATE '02000'; 
        DECLARE col_cursor CURSOR FOR SELECT COLUMN_NAME FROM test_prepare_vw;
        DECLARE pri_cursor CURSOR FOR SELECT COLUMN_NAME FROM test_prepare_vw2;
        DECLARE upd_cursor CURSOR FOR SELECT COLUMN_NAME FROM test_prepare_vw3;
        DECLARE rec_cursor CURSOR FOR SELECT COLUMN_NAME FROM test_prepare_vw4;
        DECLARE CONTINUE HANDLER FOR cursor_end SET done = 1; 

        DROP VIEW IF EXISTS test_prepare_vw; 
        DROP VIEW IF EXISTS test_prepare_vw2; 
        DROP VIEW IF EXISTS test_prepare_vw3; 
        DROP VIEW IF EXISTS test_prepare_vw4; 

        SET u_trigger = '';
        SET u_trigger = CONCAT('DELIMITER $$ \nDROP TRIGGER IF EXISTS tra_',s_tablename,'_update;\n');
        SET u_trigger = CONCAT(u_trigger,'CREATE TRIGGER tra_',s_tablename,'_update AFTER UPDATE ON ',s_tablename,' FOR EACH ROW \n');
        SET u_trigger = CONCAT(u_trigger,'BEGIN \n');
        SET u_trigger = CONCAT(u_trigger,'DECLARE msg VARCHAR(255); \n');

        SET d_trigger = '';
        SET d_trigger = CONCAT('DELIMITER $$ \nDROP TRIGGER IF EXISTS tra_',s_tablename,'_delete;\n');
        SET d_trigger = CONCAT(d_trigger,'CREATE TRIGGER tra_',s_tablename,'_delete AFTER DELETE ON ',s_tablename,' FOR EACH ROW \n');
        SET d_trigger = CONCAT(d_trigger,'BEGIN \n');

        SET @query = CONCAT('CREATE VIEW test_prepare_vw2 as SELECT COLUMN_NAME FROM INFORMATION_SCHEMA.COLUMNS WHERE table_name = \'', s_tablename, '\' AND table_schema = \'db_diers\' AND COLUMN_NAME NOT LIKE \'%updated%\' AND COLUMN_KEY = \'PRI\' ORDER BY ORDINAL_POSITION'); 
        PREPARE stmt from @query; 
        EXECUTE stmt; 
        DEALLOCATE PREPARE stmt; 

        OPEN pri_cursor;
        FETCH pri_cursor INTO s_key; 
        CLOSE pri_cursor; 
        DROP VIEW test_prepare_vw2; 

        SET @query = CONCAT('CREATE VIEW test_prepare_vw3 as SELECT COLUMN_NAME FROM INFORMATION_SCHEMA.COLUMNS WHERE table_name = \'', s_tablename, '\' AND table_schema = \'db_diers\' AND COLUMN_NAME LIKE \'%updatedby%\' AND COLUMN_KEY <> \'PRI\' ORDER BY ORDINAL_POSITION'); 
        PREPARE stmt from @query; 
        EXECUTE stmt; 
        DEALLOCATE PREPARE stmt; 

        OPEN upd_cursor;
        FETCH upd_cursor INTO s_updatedby; 
        CLOSE upd_cursor; 
        DROP VIEW test_prepare_vw3; 
        SET s_updated = LEFT(s_updatedby,(LENGTH(RTRIM(s_updatedby)))-2);

        SET @query = CONCAT('CREATE VIEW test_prepare_vw4 as SELECT COLUMN_NAME FROM INFORMATION_SCHEMA.COLUMNS WHERE table_name = \'', s_tablename, '\' AND table_schema = \'db_diers\' AND COLUMN_NAME LIKE \'%record_version%\' AND COLUMN_KEY <> \'PRI\' ORDER BY ORDINAL_POSITION'); 
        PREPARE stmt from @query; 
        EXECUTE stmt; 
        DEALLOCATE PREPARE stmt; 

        OPEN rec_cursor;
        FETCH rec_cursor INTO s_recversion; 
        CLOSE rec_cursor; 
        DROP VIEW test_prepare_vw4; 

        SET @query = CONCAT('CREATE VIEW test_prepare_vw as SELECT COLUMN_NAME FROM INFORMATION_SCHEMA.COLUMNS WHERE table_name = \'', s_tablename, '\' AND table_schema = \'db_diers\' AND COLUMN_KEY <> \'PRI\' ORDER BY ORDINAL_POSITION'); 
        PREPARE stmt from @query; 
        EXECUTE stmt; 
        DEALLOCATE PREPARE stmt; 

        SET u_trigger = CONCAT(u_trigger,'   IF (ISNULL(NEW.',s_recversion,') OR OLD.',s_recversion,' >= NEW.',s_recversion,' OR ISNULL(NEW.',s_updatedby,') OR NEW.',s_updatedby,' = \'\' OR ISNULL(NEW.',s_updated,') OR NEW.',s_updated,' = OLD.',s_updated,') THEN \n');
        SET u_trigger = CONCAT(u_trigger,'      set msg = \'Cannot update record without specifying updated/updatedby by columns and without incrementing the record version.\'; \n');
        SET u_trigger = CONCAT(u_trigger,'      SIGNAL SQLSTATE \'45000\' SET MESSAGE_TEXT = msg; \n');
        SET u_trigger = CONCAT(u_trigger,'   END IF;     \n');

        OPEN col_cursor;

        FETCH col_cursor INTO s_fieldname; 
        WHILE done = 0 DO 
            SET u_trigger = CONCAT(u_trigger,'   IF (IFNULL(OLD.',s_fieldname,',\'\') <> IFNULL(NEW.',s_fieldname,',\'\') ) THEN\n');
            SET u_trigger = CONCAT(u_trigger,'     INSERT INTO tb_auditdetail (type, tablename, pk, fieldname, oldvalue, newvalue, updatedate, username, dbusername, machinename) \n');
            SET u_trigger = CONCAT(u_trigger,'     VALUES (\'U\', \'',s_tablename,'\', OLD.',s_key,', \'',s_fieldname,'\', OLD.',s_fieldname,', NEW.',s_fieldname,', CURRENT_TIMESTAMP,NEW.',s_updatedby,',CURRENT_USER(),@@hostname);\n');
            SET u_trigger = CONCAT(u_trigger,'   END IF;\n'); 

            SET d_trigger = CONCAT(d_trigger,'     INSERT INTO tb_auditdetail (type, tablename, pk, fieldname, oldvalue, newvalue, updatedate, username, dbusername, machinename) \n');
            SET d_trigger = CONCAT(d_trigger,'     VALUES (\'D\', \'',s_tablename,'\', OLD.',s_key,', \'',s_fieldname,'\', OLD.',s_fieldname,',NULL, CURRENT_TIMESTAMP,NULL,CURRENT_USER(),@@hostname);\n');

            FETCH col_cursor INTO s_fieldname; 
        END WHILE; 
        CLOSE col_cursor; 

        DROP VIEW test_prepare_vw; 

        SET u_trigger = CONCAT(u_trigger,'END;$$ \nDELIMITER ; \n');
        SET d_trigger = CONCAT(d_trigger,'END;$$ \nDELIMITER ; \n');
        SELECT u_trigger INTO u_trigger_out;
        SELECT d_trigger INTO d_trigger_out;


    END; $$

    DELIMITER ;

    -- And finally, to extract the Trigger Scripts

    call sp_maketrigger('tb_company',@s_line1,@d_line1);

    SELECT CONCAT(@s_line1,@d_line1)

    -- You just need to copy, paste and execute the trigger script, and
    -- voila, your audit is in place.

上面的示例理所当然地认为，每个表都有 5 列：created、createdby、updated、updatedby、record_version。

但是，您可以根据需要以不同方式自定义存储过程 sp_maketrigger。 sp 也有待增强和改进。

【讨论】：

【参考方案5】：

数据库表

假设我们有一个包含以下两个表的库应用程序：

存储新旧行状态的最佳方式是使用 JSON 列。因此，对于您要启用审核日志记录的每个表，您可以创建一个审核日志表，如下所示：

CREATE TABLE book_audit_log (
    book_id BIGINT NOT NULL, 
    old_row_data JSON,
    new_row_data JSON,
    dml_type ENUM('INSERT', 'UPDATE', 'DELETE') NOT NULL,
    dml_timestamp TIMESTAMP NOT NULL,
    dml_created_by VARCHAR(255) NOT NULL,
    PRIMARY KEY (book_id, dml_type, dml_timestamp)
)

book_id 列存储已创建、更新或删除的 book 行的标识符。 old_row_data 是一个 JSON 列，它将在执行 INSERT、UPDATE 或 DELETE 语句之前捕获 book 记录的状态。 new_row_data 是一个 JSON 列，将在执行 INSERT、UPDATE 或 DELETE 语句后捕获book 记录的状态。 dml_type 是一个枚举列，用于存储创建、更新或删除给定 book 记录的 DML 语句类型。 dml_timestamp 存储 DML 语句执行时间戳。 dml_created_by 存储发出 INSERT、UPDATE 或 DELETE DML 语句的应用程序用户。

使用触发器拦截 INSERT、UPDATE 和 DELETE DML 语句

现在，要输入审计日志表，您需要创建以下 3 个触发器：

CREATE TRIGGER book_insert_audit_trigger
AFTER INSERT ON book FOR EACH ROW 
BEGIN
    INSERT INTO book_audit_log (
        book_id,
        old_row_data,
        new_row_data,
        dml_type,
        dml_timestamp,
        dml_created_by
    )
    VALUES(
        NEW.id,
        null,
        JSON_OBJECT(
            "title", NEW.title,
            "author", NEW.author,
            "price_in_cents", NEW.price_in_cents,
            "publisher", NEW.publisher
        ),
        'INSERT',
        CURRENT_TIMESTAMP,
        @logged_user
    );
END

CREATE TRIGGER book_update_audit_trigger
AFTER UPDATE ON book FOR EACH ROW 
BEGIN
    INSERT INTO book_audit_log (
        book_id,
        old_row_data,
        new_row_data,
        dml_type,
        dml_timestamp,
        dml_created_by
    )
    VALUES(
        NEW.id,
        JSON_OBJECT(
            "title", OLD.title,
            "author", OLD.author,
            "price_in_cents", OLD.price_in_cents,
            "publisher", OLD.publisher
        ),
        JSON_OBJECT(
            "title", NEW.title,
            "author", NEW.author,
            "price_in_cents", NEW.price_in_cents,
            "publisher", NEW.publisher
        ),
        'UPDATE',
        CURRENT_TIMESTAMP,
        @logged_user
    );
END

CREATE TRIGGER book_delete_audit_trigger
AFTER DELETE ON book FOR EACH ROW 
BEGIN
    INSERT INTO book_audit_log (
        book_id,
        old_row_data,
        new_row_data,
        dml_type,
        dml_timestamp,
        dml_created_by
    )
    VALUES(
        OLD.id,
        JSON_OBJECT(
            "title", OLD.title,
            "author", OLD.author,
            "price_in_cents", OLD.price_in_cents,
            "publisher", OLD.publisher
        ),
        null,
        'DELETE',
        CURRENT_TIMESTAMP,
        @logged_user
    );
END

JSON_OBJECT MySQL 函数允许我们创建一个 JSON 对象，该对象采用提供的键值对。

dml_type 列的值设置为INSERT、UPDATE 或DELETE，dml_timestamp 的值设置为CURRENT_TIMESTAMP。

dml_created_by 列设置为@logged_user MySQL 会话变量的值，该变量先前由应用程序使用当前登录的用户设置：

Session session = entityManager.unwrap(Session.class);

Dialect dialect = session.getSessionFactory()
    .unwrap(SessionFactoryImplementor.class)
    .getJdbcServices()
    .getDialect();

session.doWork(connection -> 
    update(
        connection,
        String.format(
            "SET @logged_user = '%s'", 
            ReflectionUtils.invokeMethod(
                dialect,
                "escapeLiteral",
                LoggedUser.get()
            )
        )
    );
);

测试时间

在book 表上执行 INSERT 语句时：

INSERT INTO book (
    id,
    author, 
    price_in_cents, 
    publisher, 
    title
) 
VALUES (
    1,
    'Vlad Mihalcea', 
    3990, 
    'Amazon', 
    'High-Performance Java Persistence 1st edition'
)

我们可以看到book_audit_log中插入了一条记录，该记录捕获了刚刚在book表上执行的INSERT语句：

| book_id | old_row_data | new_row_data                                                                                                                         | dml_type | dml_timestamp       | dml_created_by |
|---------|--------------|--------------------------------------------------------------------------------------------------------------------------------------|----------|---------------------|----------------|
| 1       |              | "title": "High-Performance Java Persistence 1st edition", "author": "Vlad Mihalcea", "publisher": "Amazon", "price_in_cents": 3990 | INSERT   | 2020-07-29 13:40:15 | Vlad Mihalcea  |

更新book 表行时：

UPDATE book 
SET price_in_cents = 4499 
WHERE id = 1

我们可以看到book 表上的AFTER UPDATE 触发器将向book_audit_log 添加一条新记录：

| book_id | old_row_data                                                                                                                         | new_row_data                                                                                                                         | dml_type | dml_timestamp       | dml_created_by |
|---------|--------------------------------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------|----------|---------------------|----------------|
| 1       |                                                                                                                                      | "title": "High-Performance Java Persistence 1st edition", "author": "Vlad Mihalcea", "publisher": "Amazon", "price_in_cents": 3990 | INSERT   | 2020-07-29 13:40:15 | Vlad Mihalcea  |
| 1       | "title": "High-Performance Java Persistence 1st edition", "author": "Vlad Mihalcea", "publisher": "Amazon", "price_in_cents": 3990 | "title": "High-Performance Java Persistence 1st edition", "author": "Vlad Mihalcea", "publisher": "Amazon", "price_in_cents": 4499 | UPDATE   | 2020-07-29 13:50:48 | Vlad Mihalcea  |

删除book 表行时：

DELETE FROM book 
WHERE id = 1

book 表上的 AFTER DELETE 触发器将新记录添加到 book_audit_log：

| book_id | old_row_data                                                                                                                         | new_row_data                                                                                                                         | dml_type | dml_timestamp       | dml_created_by |
|---------|--------------------------------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------|----------|---------------------|----------------|
| 1       |                                                                                                                                      | "title": "High-Performance Java Persistence 1st edition", "author": "Vlad Mihalcea", "publisher": "Amazon", "price_in_cents": 3990 | INSERT   | 2020-07-29 13:40:15 | Vlad Mihalcea  |
| 1       | "title": "High-Performance Java Persistence 1st edition", "author": "Vlad Mihalcea", "publisher": "Amazon", "price_in_cents": 3990 | "title": "High-Performance Java Persistence 1st edition", "author": "Vlad Mihalcea", "publisher": "Amazon", "price_in_cents": 4499 | UPDATE   | 2020-07-29 13:50:48 | Vlad Mihalcea  |
| 1       | "title": "High-Performance Java Persistence 1st edition", "author": "Vlad Mihalcea", "publisher": "Amazon", "price_in_cents": 4499 |                                                                                                                                      | DELETE   | 2020-07-29 14:05:33 | Vlad Mihalcea  |

就是这样！

【讨论】：

以上是关于如何创建触发器以将更改事件添加到审核日志表中的主要内容，如果未能解决你的问题，请参考以下文章

创建触发器以将数据从一个表获取到另一个表并生成时间戳

组合框更改事件正在触发工作表中的任何更改

在 Mysql 中添加新列时如何轻松维护审计触发器

如何创建触发器以将值插入到插入时为 Max([ID Field])+1 的 ID 字段中

您将如何审核 ASP.NET 成员表，同时记录哪些用户进行了更改？

在oracle中插入表B后创建触发器以将新行添加到表A中时出错