Yet another DDL dump script
Inspired by some recent posts about scripts for parsing mysqldump files, I thought I’d put my own little dump program out there for those looking for alternatives to mysqldump. It’s written in perl and only actually uses the mysqldump utility to dump the data portion of its file dumps. The rest is done via mysql’s internal show commands. Each database is dumped to it’s own directory, where each component in that database is dumped into five subdirs: schemas, routines, views, triggers & data. Inside these subdirs is a file per table and file per proc, etc. The data section is slightly different: as I mentioned it uses mysqldump to dump bulk insert statements into a sql file, one per table. So at any time, any or all components (including data) can be reloaded with a simple redirect or pipe back to mysql client:
[user@svr123]$ mysql -D dbname < /path/to/files/dbname/schemas/*.sql (data files are zipped)
[user@svr123]$ mysql -D dbname < /path/to/files/dbname/routines/*.sql
[user@svr123]$ mysql -D dbname < /path/to/files/dbname/views/*.sql
[user@svr123]$ mysql -D dbname < /path/to/files/dbname/triggers/*.sql
[user@svr123]$ gunzip -c /path/to/files/dbname/data/*.sql.gz|mysql -D dbname
The script is self-explanatory and since I’m pretty much a newb at perl code, I’d love to know if there’s a better way to do any of this stuff. You can find the code here.