我正在将一个t-sql存储过程转换成mysql。我对t-sql不太熟悉，正在努力使自己对temp表和存储过程更加熟悉。提前谢谢你的帮助。
t-sql的原始版本如下所示（编辑：注意，这只是用于生成报表的原始过程的一部分）：

DROP TABLE IF EXISTS #accounts;

SELECT  d.data_id                   
        ,p.pp_name AS name
        ,CONVERT(tinyint,1) AS flag             
        ,d.pd_date
        ,CONVERT(char(6),pd_date,112) AS date_period                
        ,CONVERT(varchar(3),0) AS n_phones                          
INTO    #accounts                   
from    table_detail d                  
    JOIN    table_pool p ON d.pp_id = p.pp_id                   
    JOIN    table_type t ON p.pp_type_id = t.pp_type_id                                                     
    JOIN    Inventory i ON d.data_id = i.data_id                    
    JOIN    Product pr ON i.product_id = pr.product_id  
WHERE   pp_name IN (SELECT name FROM Sandbox..desired_sandbox)                  
    AND DATEDIFF(MONTH,pd_date,GETDATE()) < 3

UPDATE  a                   
SET     a.flag = 0              
FROM    #accounts a                 
JOIN    table_detail d ON a.data_id = d.data_id                 
JOIN    table_pool p ON d.pp_id = p.pp_id                   
WHERE   d.pd_date < a.pd_date                   
    AND pp_name != 'error';

我目前的更新如下。我需要把这个包起来吗 CREATE TEMPORARY TABLE IF NOT EXISTS accounts AS (<insert query here>) 而不是 INTO #accounts ?

SELECT  d.data_id                   
        ,p.pp_name AS name                              
        ,CONVERT(tinyint,1) AS flag             
        ,d.pd_date
        ,DATE_FORMAT(pd_date,'%Y%m%d') AS date_period               
        ,CONVERT(varchar(3),0) AS n_phones                                  
FROM    table_detail d                  
    JOIN    table_pool p ON d.pp_id = p.pp_id                   
    JOIN    table_type t ON p.pp_type_id = t.pp_type_id                                 
    JOIN    Inventory i ON d.data_id = i.data_id                    
    JOIN    Product pr ON i.product_id = pr.product_id  
WHERE   pp_name IN (SELECT name FROM Sandbox..desired_sandbox)              
    AND TIMESTAMPDIFF(MONTH,pd_date,NOW()) < 3

然后在语法正确的情况下执行以下操作：

UPDATE  accounts a -- Is this the correct way to add an alias and update the temp table?                
    JOIN    table_detail d ON a.data_id = d.data_id                 
    JOIN    table_pool p ON d.pp_id = p.pp_id                   
SET     a.flag = 0              
WHERE   d.pd_date < a.pd_date                   
    AND pp_name != 'error';

最后，我假设我可以按照这篇文章将最终的查询 Package 成一个存储过程，对吗？要在帖子中总结代码：

drop procedure if exists procedure_name;
DELIMITER $$
create procedure procedure_name ()
BEGIN
    DROP TEMPORARY TABLE IF EXISTS accounts;

    CREATE TEMPORARY TABLE accounts AS (
    SELECT...
    FROM...
    WHERE...
    ;
    )

    UPDATE accounts
        JOIN ...
        JOIN ...
    SET...
    WHERE...;

    DROP TEMPORARY TABLE accounts; -- otherwise it survives the stored proc call
END
$$ -- signify end of block
DELIMITER ; -- reset to default delimiter