Will Leshner <[EMAIL PROTECTED]> wrote: > In the article about locking and concurrency on the SQLite website, > where it talks about how to corrupt a SQLite database, it says that a > SQLite database can be corrupted if a hot journal file is missing > when SQLite reconnects to the database. What kind of corruption are > we talking about, exactly? Is it that the data in the database is > inconsistent, in the sense that only some of the data has been > written with no way to roll any of it back? Or is this a more serious > kind of corruption in which the database file becomes completely > unusable? I guess I'd always thought the latter, but after reading > through the article more carefully, I am now thinking it might be the > former. If it is the former, it might be useful to make that explicit > in the article. While having an inconsistent database is certainly a > BadThing, having a completely unusable database (one that SQLite > can't even open anymore) is far worse, in my opinion. >
The database might be completely unusable. It depends on which writes completed and which had not at the time of the crash. -- D. Richard Hipp <[EMAIL PROTECTED]>