对于表格:

SELECT 
  TABLE_NAME,COLUMN_NAME,CONSTRAINT_NAME, REFERENCED_TABLE_NAME,REFERENCED_COLUMN_NAME
FROM
  INFORMATION_SCHEMA.KEY_COLUMN_USAGE
WHERE
  REFERENCED_TABLE_SCHEMA = '<database>' AND
  REFERENCED_TABLE_NAME = '<table>';

对于列:

SELECT 
  TABLE_NAME,COLUMN_NAME,CONSTRAINT_NAME, REFERENCED_TABLE_NAME,REFERENCED_COLUMN_NAME
FROM
  INFORMATION_SCHEMA.KEY_COLUMN_USAGE
WHERE
  REFERENCED_TABLE_SCHEMA = '<database>' AND
  REFERENCED_TABLE_NAME = '<table>' AND
  REFERENCED_COLUMN_NAME = '<column>';

基本上，我们用where子句中的REFERENCED_COLUMN_NAME更改了REFERENCED_TABLE_NAME。

我不愿意再添加另一个答案，但我必须从其他人那里乞求、借用和窃取才能得到我想要的答案，这是给定模式中表上所有FK关系的完整列表，包括其他模式中表的FK关系。两个关键的记录集是information_schema。KEY_COLUMN_USAGE和information_schema.referential_constraints。如果缺少您想要的属性，只需取消KCU的注释。RC。看看有什么空房

SELECT DISTINCT KCU.TABLE_NAME, KCU.COLUMN_NAME, REFERENCED_TABLE_SCHEMA, KCU.REFERENCED_TABLE_NAME, KCU.REFERENCED_COLUMN_NAME, UPDATE_RULE, DELETE_RULE #, KCU.*, RC.*
FROM information_schema.KEY_COLUMN_USAGE KCU
INNER JOIN information_schema.referential_constraints RC ON KCU.CONSTRAINT_NAME = RC.CONSTRAINT_NAME
WHERE TABLE_SCHEMA = (your schema name)
AND KCU.REFERENCED_TABLE_NAME IS NOT NULL
ORDER BY KCU.TABLE_NAME, KCU.COLUMN_NAME;

我需要鸟瞰表之间的关系(在ORM中使用)。使用该页面的建议，经过试验，我整理了以下查询:

SELECT
    KCU.CONSTRAINT_NAME,
    KCU.TABLE_NAME,
    KCU.COLUMN_NAME,
    KCU.REFERENCED_TABLE_NAME,
    KCU.REFERENCED_COLUMN_NAME
FROM
    INFORMATION_SCHEMA.KEY_COLUMN_USAGE AS KCU
    JOIN INFORMATION_SCHEMA.COLUMNS AS COLS
        ON
                COLS.TABLE_SCHEMA = KCU.TABLE_SCHEMA
            AND COLS.TABLE_NAME   = KCU.TABLE_NAME
            AND COLS.COLUMN_NAME  = KCU.COLUMN_NAME
WHERE
        KCU.CONSTRAINT_SCHEMA = {YOUR_SCHEMA_NAME}
    AND KCU.REFERENCED_TABLE_NAME IS NOT NULL
ORDER BY
    KCU.TABLE_NAME,
    COLS.ORDINAL_POSITION

它只返回我需要的东西，而且是按照我想要的顺序。

我还对结果做了一些处理(将其转换为某种字典)，以便它可以用于创建聚合。

了解更新和删除行为通常是有帮助的，这是其他答案没有提供的。现在开始。

SELECT cu.table_name,
       cu.column_name,
       cu.constraint_name,
       cu.referenced_table_name,
       cu.referenced_column_name,
       IF(rc.update_rule = 'NO ACTION', 'RESTRICT', rc.update_rule) AS update_rule,-- See: https://stackoverflow.com/a/1498015/2742117
       IF(rc.delete_rule = 'NO ACTION', 'RESTRICT', rc.delete_rule) AS delete_rule -- See: https://stackoverflow.com/a/1498015/2742117
FROM information_schema.key_column_usage cu
INNER JOIN information_schema.referential_constraints rc ON rc.constraint_schema = cu.table_schema
AND rc.table_name = cu.table_name
AND rc.constraint_name = cu.constraint_name
WHERE cu.referenced_table_schema = '<your schema>'
  AND cu.referenced_table_name = '<your table>';

我提出的解决方案是脆弱的;它依赖于django的外键命名约定。

USE information_schema;
tee mysql_output
SELECT * FROM TABLE_CONSTRAINTS WHERE CONSTRAINT_TYPE = 'FOREIGN KEY' AND TABLE_SCHEMA = 'database_name';
notee

然后在壳层中，

grep 'refs_tablename_id' mysql_output

对于表格:

SELECT 
  TABLE_NAME,COLUMN_NAME,CONSTRAINT_NAME, REFERENCED_TABLE_NAME,REFERENCED_COLUMN_NAME
FROM
  INFORMATION_SCHEMA.KEY_COLUMN_USAGE
WHERE
  REFERENCED_TABLE_SCHEMA = '<database>' AND
  REFERENCED_TABLE_NAME = '<table>';

对于列:

SELECT 
  TABLE_NAME,COLUMN_NAME,CONSTRAINT_NAME, REFERENCED_TABLE_NAME,REFERENCED_COLUMN_NAME
FROM
  INFORMATION_SCHEMA.KEY_COLUMN_USAGE
WHERE
  REFERENCED_TABLE_SCHEMA = '<database>' AND
  REFERENCED_TABLE_NAME = '<table>' AND
  REFERENCED_COLUMN_NAME = '<column>';

基本上，我们用where子句中的REFERENCED_COLUMN_NAME更改了REFERENCED_TABLE_NAME。