Question

I am writing an iPhone application where the user receives multiple messages from different users. These messages are stored in an sqlite3 database. With time the user might like to delete received messages from one user, but for sure he will continue to receive new messages from that user after deleting the old ones.

Since retrieving the messages will be done using a SELECT statement, which scenario is better to use when the user would like to delete the messages (in terms of performance):

1. DELETE all the old messages normally and continue to retrieve the new ones using a statement like: SELECT Messages FROM TableName WHERE UserID = (?)
2. Add a field to the table of type INTEGER and upon the DELETE request set this field to 1 and after that retrieve the new messages using a statement like: SELECT Messages FROM TableName WHERE UserID = (?) AND IsDeleted = 0

One more thing, if scenario 1 is used (normal DELETE) will this cause any fragmentation of the database file on the disk?

Many thanks in advance.

Answer 1

Using scenario 1 is much better, since both SELECT and DELETE in SQL operate at the same level of speed and scenario 1 will grant you not having dangling tuples (Unwanted Rows) in your database.

If you are wishing to perform data backup after any deletion process so scenario 2 is a must but you have to take into consideration the growing size of your database which leads to a slower performance in future.

Finally I would like to add that performing deleting operations on a database would not cause any fragmentation issues since most of databases have fragmentation and optimizing tools in their engines.