On 25 Nov 2010, at 5:25pm, Tito Ciuro wrote: > 1) SELECT on sqlite_master. All SQLite3 databases contain it, so it would > fail right away is it wasn't a valid database. > 2) Issue a "PRAGMA integrity_check" statement, if (1) succeeds. > > However, I'm wondering whether (1) is redundant. If the integrity check > detects right away that the file is not a SQLite3 database, then we're good. > But if the integrity check reads the whole file before failing, perhaps > running (1) would not be such a bad after all. > > Also, if there is a need to know whether a database is corrupted, an > optimization seems to be to issue "PRAGMA integrity_check (1)" statement, > which would bail out at the first sign of trouble. As stated in the docs: > http://www.sqlite.org/pragma.html
I believe that if Nick's going to do (2) anyway, (1) is redundant. In my own case I couldn't do (2) because some of the databases I was dealing with would have made it take too long. And my hardware were extremely trustworthy anyway: I was really just checking that the files were the right files, and someone hadn't (for example) renamed a journal file as the database file WHICH IS SOMETHING SOMEONE ACTUALLY DID TO ME ONCE ARGHHHHH ! And I agree Nick only needs to see the first error from 'PRAGMA integrity_check' so he can use 1 as a parameter. Simon. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
