I have an application which collects data into a mysql table. The table has no unique id column, so I can't reference a specific row by id.
I want to write a dump application which every day dumps the new rows added to the table to upload them elsewhere. I could do it by adding a unique id field and storing the last id dumped, but I don't want to add an id column just for that to the table.
So I thought I store the number of rows in the table at every dump and use that number as an offset next time the table is dumped (select * from table limit verylargenumber offset x). Of course, it works only if there is a guarantee new rows always inserted at the end of the table, so all new rows will be after the offset.
I assume I can rely on that. Am I right?