Limiting the number of records from mysqldump?

As skaffman says, use the –where option:

mysqldump --opt --where="1 limit 1000000" database

Of course, that would give you the first million rows from every table.