Igor Tandetnik <itandet...@mvps.org> wrote: > > > Is there a reason that sqlite3_open() will happily "open" a non-sqlite > > file, returning SQLITE_OK, instead of returning SQLITE_NOTADB, which
> > would seem a more obvious return value? > > SQLite doesn't actually touch the file until the first "substantive" statement is executed on the connection. This allows one to set various PRAGMAs that can only be set before the database is > created. Ah, that makes sense, ok. > > It will also happily open a > > sqlite database that it will later say is corrupt when I try to query > > some data, implying it doesn't do data integrity checks on the > > database on opening? > > Do you really want SQLite to scan a gigabytes-large file every time it's opened? Depends - obviously not if you need quick access to the data, however if the gigabytes-large file is corrupt, do you really want to wait until you execute the first "substantive" statement to discover the DB is corrupt? It makes coding for corrupt database more complex if you only find out (possibly much) further down the code stream... > > > So... Assuming the above is how sqlite is meant to work, what would be > > the best way to ascertain if a successfully opened file is ok to use? > > Run "PRAGMA integrity_check" right after opening. Ah, that's what I was looking for! Thanks Igor, I'll try that. Thanks, Nick. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
