On Wed, May 19, 2010 at 5:53 PM, Jonathan Duncan <[email protected]> wrote: > I think I like this way the best: > http://www.everymanhosting.com/forum/about22.html > > However, this will not address the required cascading update. > > This has some good information: > http://blog.mclaughlinsoftware.com/2009/05/25/mysql-merge-gone-awry/ > > But also not a final solution. > > I am thinking I may just write a PHP script that will select from db1.table1, > insert it into db2.table1 and then check if there is an associated row in > db1.table2 and if so insert that in db2.table 2 with new 'id' key from last > insert on db2.table1.
You will need a PHP script for sure. This is one problem where an incrementing primary key causes pain. Your first article assumes you have no foreign key dependencies, or possibly a row-dependent key like a username. The second article does too. If auto_increment primary keys weren't so simple, easy, and beautiful, they would be banished to hell for problems like this. A UUID approach to generated keys would be more robust, but adds too much complexity. And really, how often do these types of table merges happen? _______________________________________________ UPHPU mailing list [email protected] http://uphpu.org/mailman/listinfo/uphpu IRC: #uphpu on irc.freenode.net
