Question

我正在尝试使用游标来处理包含字符串的行：

CREATE PROCEDURE REVERT_ALL(IN TABLE_NAME VARCHAR(255))
  BEGIN
    DECLARE bDone INT;
    DECLARE CH_ID INT;
    DECLARE CH_CHANGE CHAR;

    DECLARE curs cursor for 
      SELECT `table_id`, `change_type` FROM mysql_snapshot.db_changes where `table_name` = "rooms";
    DECLARE CONTINUE HANDLER FOR NOT FOUND SET bDone = 1;

    OPEN curs;

    insert into splog set text = concat('SELECT id, `table_name`, table_id, `change_type` FROM mysql_snapshot.db_changes where table_name = ',TABLE_NAME, ';');

    SET bDone = 0;
    REPEAT
      FETCH curs INTO CH_ID, CH_CHANGE;

      insert into splog set text = concat_ws( ' -- ', 'CH_ID ', CH_ID,' TABLE ',TABLE_NAME, ' CH_CHANGE ', CH_CHANGE);

      UNTIL bDone END REPEAT;

      CLOSE curs;
    END;

由于我正在搜索大部分时间的原因，

where `table_name` = "rooms"

似乎被忽略了。计划是将其改为

where `table_name` = TABLE_NAME

使用procedure参数。我只是获得所有行。插件用于记录和调试。

Answer 1

您在过程参数名称TABLE_NAME和表列名table_name之间存在歧义。应该避免它，因为它会导致像这样的模糊问题。

在这种情况下，在该过程中，反引号中的TABLE_NAME，table_name和table_name都是interpreted as the local variable name（参数名称）。所以，你的条件'where table_name = ',TABLE_NAME总是如此;当您使用TABLE_NAME文字替换"rooms"时，如果您使用"rooms"作为参数调用过程，则条件始终为true，否则始终为false。考虑这个简化的例子：

DROP PROCEDURE IF EXISTS pr;
DROP TABLE IF EXISTS t;
CREATE PROCEDURE pr (IN TABLE_NAME VARCHAR(255)) 
  SELECT `id`, `table_name`, TABLE_NAME FROM t
;

CREATE TABLE t (id INT, `table_name` VARCHAR(255));
INSERT INTO t VALUES (1,'hotels'),(2,'rooms');

CALL pr("rooms");
CALL pr("foo");

你会得到

MariaDB [test]> CALL pr("rooms");
+------+--------------+------------+
| id   | `table_name` | TABLE_NAME |
+------+--------------+------------+
|    1 | rooms        | rooms      |
|    2 | rooms        | rooms      |
+------+--------------+------------+
2 rows in set (0.00 sec)

MariaDB [test]> CALL pr("foo");
+------+--------------+------------+
| id   | `table_name` | TABLE_NAME |
+------+--------------+------------+
|    1 | foo          | foo        |
|    2 | foo          | foo        |
+------+--------------+------------+
2 rows in set (0.00 sec)

如您所见，无论参数如何，对于所有行，table_name始终等于TABLE_NAME。

现在，如果您使用显式TABLE_NAME替换"rooms"：

DROP PROCEDURE IF EXISTS pr;
DROP TABLE IF EXISTS t;
CREATE PROCEDURE pr (IN TABLE_NAME VARCHAR(255)) 
  SELECT `id`, `table_name`, "rooms" FROM t
;

CREATE TABLE t (id INT, `table_name` VARCHAR(255));
INSERT INTO t VALUES (1,'hotels'),(2,'rooms');

CALL pr("rooms");
CALL pr("foo");

现在，对于所有行，条件table_name = "rooms"对于所有行都是正确的，对于第二次调用，对于所有行，条件MariaDB [test]> CALL pr("rooms"); +------+--------------+-------+ | id | `table_name` | rooms | +------+--------------+-------+ | 1 | rooms | rooms | | 2 | rooms | rooms | +------+--------------+-------+ 2 rows in set (0.00 sec) Query OK, 0 rows affected (0.00 sec) MariaDB [test]> CALL pr("foo"); +------+--------------+-------+ | id | `table_name` | rooms | +------+--------------+-------+ | 1 | foo | rooms | | 2 | foo | rooms | +------+--------------+-------+ 2 rows in set (0.00 sec)将为真：

DROP PROCEDURE IF EXISTS pr;
DROP TABLE IF EXISTS t;
CREATE PROCEDURE pr (IN T_NAME VARCHAR(255)) 
  SELECT `id`, `table_name`, T_NAME FROM t
;

CREATE TABLE t (id INT, `table_name` VARCHAR(255));
INSERT INTO t VALUES (1,'hotels'),(2,'rooms');

CALL pr("rooms");
CALL pr("foo");

您需要的是参数的不同名称：

table_name = T_NAME

现在比较MariaDB [test]> CALL pr("rooms"); +------+------------+--------+ | id | table_name | T_NAME | +------+------------+--------+ | 1 | hotels | rooms | | 2 | rooms | rooms | +------+------------+--------+ 2 rows in set (0.00 sec) Query OK, 0 rows affected (0.00 sec) MariaDB [test]> CALL pr("foo"); +------+------------+--------+ | id | table_name | T_NAME | +------+------------+--------+ | 1 | hotels | foo | | 2 | rooms | foo | +------+------------+--------+ 2 rows in set (0.00 sec)实际上是有道理的;

use unique_type_id::UniqueTypeId;
#[derive(UniqueTypeId)]
struct Test1;
#[derive(UniqueTypeId)]
struct Test2;

assert_eq!(Test1::id().0, 1u64);
assert_eq!(Test2::id().0, 2u64);

在mysql游标中似乎被忽略的地方

1 个答案: