我无法弄清楚为什么在mysql proc之外运行的完全相同的查询不会在其中返回相同的数据。
这是proc。
DELIMITER $$
USE `poll_app`$$
DROP PROCEDURE IF EXISTS `record_new_vote`$$
CREATE DEFINER=`somemysqluser`@`%` PROCEDURE `record_new_vote`(
p_member_id INT (10),
p_option_ids VARCHAR(255), -- 12,13,14,15
p_option_id INT (10),
p_stream_id INT (10),
p_scene_id INT (10)
)
BEGIN
DECLARE _user_already_voted VARCHAR(255);
DECLARE _update_off_id INT;
DECLARE _update_on_id INT;
-- 1. Query to figure out what to update
SELECT
p_member_id, p_option_ids, p_option_id, p_stream_id, p_scene_id;
SET SESSION group_concat_max_len = 1000000;
SELECT
GROUP_CONCAT(selections.status, "~~~", selections.id, "~~~", selections.option_id),
GROUP_CONCAT(IF(selections.status = "ON", id, NULL)),
GROUP_CONCAT(IF(selections.status = "OFF" && selections.option_id = p_option_id, selections.id, NULL))
INTO
_user_already_voted, _update_off_id, _update_on_id
FROM
selections
WHERE
selections.member_id = p_member_id
AND
selections.option_id IN (p_option_ids)
AND
selections.stream_id = p_stream_id
AND
selections.scene_id = p_scene_id;
-- 2. Check to see if the user has already voted on the poll or not
IF (_user_already_voted IS NOT NULL) THEN
SELECT
_user_already_voted, _update_off_id, _update_on_id;
END IF;
END$$
DELIMITER ;
这是输出的内容
[ [ RowDataPacket {
p_member_id: 107,
p_option_ids: '1005,1006,1007,1008',
p_option_id: 1007,
p_stream_id: 7,
p_scene_id: 1 } ],
[ RowDataPacket {
_user_already_voted: 'OFF~~~451~~~1005',
_update_off_id: NULL,
_update_on_id: NULL } ],
当我在proc之外运行select查询时(插入了相同的数据)
SELECT
GROUP_CONCAT(`status`, "~~~", id, "~~~", option_id) AS "selections",
GROUP_CONCAT(IF(`status` = "ON", id, NULL)) AS "id_to_turn_status_off",
GROUP_CONCAT(IF(`status` = "OFF" && option_id = 1007, id, NULL)) AS "id_to_turn_status_on"
FROM
selections
WHERE
member_id = 107
AND
option_id IN (1005,1006,1007,1008)
AND
stream_id = 7
AND
scene_id = 1
结果
selections id_to_turn_status_off id_to_turnstatus_on
OFF~~~451~~~1005,ON~~~452~~~1006 452 null
如您所见,组concat字符串不长,id_to_turn_status_off为452,而同一查询在proc中运行时变量_update_off_id为NULL。
我无法想象我的生活,任何帮助都会非常感激。
答案 0 :(得分:1)
显然
IN("1005,1006,1007,1008")
不等同于
IN(1005,1006,1007,1008)
在我在proc中的查询中,因为它接受的参数是一个字符串,它使用双引号版本" 1005,1006,1007,1008",其中我在外面运行的查询是非双引号版本1005,1006,1007,1008
这就是导致差异的原因。
我通过使用此更正
来纠正它FIND_IN_SET(selections.option_id, p_option_ids)
或在查询之外
FIND_IN_SET(selections.option_id, "1005,1006,1007,1008")
答案 1 :(得分:0)
很高兴你明白了:)
它似乎部分工作的原因是因为类型转换。
p_option_ids VARCHAR(255)
p_option_id INT (10),
selections.option_id IN (p_option_ids)
这是试图在VARCHAR中找到一个INT。 MySQL将字符串动态地转换为int,因此它会向下移动字符串,直到找到非数字字符并截断其余字符。 “1005,1006,1007,1008”转换为1005(截断“,1006,1007,1008”)。这就是它找到option_id = 1005行而不是option_id = 1007行的原因。