I'm reading very detailed tutorials on how to use transactions with database types and database engines, but I haven't found a guide that teaches me when and why I should use them.
I know transactions are usually used for banking, so when we work with money data, but I can imagine they are used in many other ways.
Today I'm working on a page with various INSERT
statements for a relational database, and I wanted to know if this is one of the cases when I should use them.
I get an impression that I don't know the cases when the data can be partially lost (apart from coder errors) so I'm always worried about when I should use them.
Can someone explain or give some link with these fundamental rules?
I'm using MySQL 5.0.8
. Should I use InnoDB
for all tables that need transactions? If yes, is InnoDB
slower than the common MyISAM
but I shouldn't worry about that?
thanks
Transactions should be used when there is the possibility that either failure to complete or someone else reading or writing in the middle of your task could cause damage to the data. These include but are not limited to:
In addition to what Nick Craver wrote, you would want to use a transaction when you have a series of writes that need to be performed atomically; that is, they should all succeed or none should succeed.
You use transactions when you have a group of actions that must be atomic (either all succeed or none succeed) Wrapping these actions in a transaction allows you to rollback actions that have already succeeded when you encounter an error. It also ensures that the data you are working with is consistent as locks will be held until the transaction is complete.
In some frameworks, e.g. Spring, automatic transactions allow to re-execute a transaction if it failed.
Basically any time you have a unit of work that is either sensitive to outside changes or needs the ability to rollback every change, if an error occurs or some other reason.
Look here for some excellent answers and their reasons for using them.