|
|
Posted by Gordon Burditt on 07/28/07 16:43
>[code=php]
>$result = mysql_query('sql query string');
>if (!$result) echo mysql_error();
>[/code]
>
>If it fails, it will display the succeeded error.
>
>Now that I gave an example, here are my doubts:
>
>1. Should this procedure be done every time a query is done, or,
>should we assume that 'SELECT' queries are always done with success if
>sql syntax is correct?
It is always possible for a MySQL query to fail due to network
errors (even if the DB is on the local host), disk errors, permission
problems, etc. Also, it's not only SYNTAX that can cause errors
in a sql query. If the table no longer matches the query (table
doesn't exist, field doesn't exist, etc.) you'll get an error.
Note that returning 0 rows in response to a select (or affecting
0 rows for a delete) is *NOT* an error.
>2. If we have a lot of queries that depend on each other (See the next
>example), imagine that query1 and query2 are made with success, but
>query3 fails, I will have inconsistent data since, the update is made
>on query2, but no delete will be made on query3 since it fails.
>How can we avoid this?
You can't. It is always possible for someone to unplug a network cable
or take down the server between queries. But you can use transactions
to be sure that all or none of the queries are committed. Check your
queries and use ROLLBACK if something fails.
>[code=php]
>$result1 = mysql_query('SELECT ...');
>if (!$result1) {
> echo mysql_error();
> exit(0);
>}
>
>$result2 = mysql_query('UPDATE ... with data supplied from $result1');
>if (!$result2) {
> echo mysql_error();
> exit(0);
>}
>
>$result3 = mysql_query('INSERT / UPDATE / DELETE ... with data
>supplied from $result1');
>if (!$result3) {
> echo mysql_error();
> exit(0);
>}
>[/code]
>
>Hope you guys can understand my examples, if not, I'll try to explain
>in a better way.
>
>Thanks for your attention and spent time on reading this.
>
Navigation:
[Reply to this message]
|