[Phpmyadmin-devel] Simulate UPDATE query

[Marc Delisle]

Hugues Peccatte a écrit :
> 2014-06-13 19:00 GMT+02:00 Ashutosh Dhundhara <ashutoshdhundhara at yahoo.com>:
> 
>> Hi,
>> I was working on RFE #861
>> <https://sourceforge.net/p/phpmyadmin/feature-requests/861/> (Simulate
>> UPDATE query).
>> Lets say I have two tables:
>>
>> `table_1`
>> +----+----------+
>> | id | value    |
>> +----+----------+
>> | 10 | value_10 |
>> | 20 | value_20 |
>> +----+----------+
>>
>> `table_2`
>> +----+----------+
>> | id | value    |
>> +----+----------+
>> | 10 | value_10 |
>> | 20 | value_20 |
>> +----+----------+
>>
>> UPDATE Query: UPDATE table_1, table_2 SET table_1.value = 'NEW_VALUE',
>> table_2.value='NEW_VALUE' WHERE table_1.id > 10 AND table_2.id > 10;
>>
>> This will affect 2 rows.
>>
>> How to simulate this query using SELECT statement?
>>
>> I was trying:
>>
>> SELECT DISTINCT table_1.value, table_2.value FROM table_1, table_2 WHERE
>> table_1.value <> 'NEW_VALUE' AND table_2.value <> 'NEW_VAUE' AND
>> table_1.id > 10 AND table_2.id > 10;
>>
>> but this only returns 1 row.
>> Can this be done in a single query only?
>>
>> Regards,
>> Ashutosh Dhundhara
>>
> 
> Hi,
> 
> Isn't it possible to run the query with
> set AUTOCOMMIT=0;
> and a rollback at the end?
> 
> Hugues.

Thanks Hugues. In the "simulate query" subtask [0] for this student, one 
  of the feature requests is [1] which involves rolling back (but thanks 
for the mention of AUTOCOMMIT=0). But this is only for transactional 
storage engines.

Another aspect of this subtask is [2] which involves a simulation via 
SELECT.

[0] 
https://wiki.phpmyadmin.net/pma/GSoC_2014_Ideas_List#Interface_improvements

[1] http://sourceforge.net/p/phpmyadmin/feature-requests/869/

[2] http://sourceforge.net/p/phpmyadmin/feature-requests/861/

-- 
Marc Delisle (phpMyAdmin)