Thanks a lot both Igor and Simon for the quick response! Your two answers settle my doubts.
On Mon, Nov 16, 2015 at 3:48 PM, Simon Slavin <slavins at bigfraud.org> wrote: > On 16 Nov 2015, at 12:24am, Randy Eels <randyeels at gmail.com> wrote: > > > My main question is: is an 'INTEGER PRIMARY KEY' column guaranteed not to > > change, except as a consequence of ordinary UPDATE statements? (And in > > particular, not as a consequence of a 'vacuum' statement.) > > Values in the rowid column (which has a number of other names) can change > only if that columns is not aliased by the table definition. Declaring a > named column as 'INTEGER PRIMARY KEY' counts as an alias. > > > I'd also thank further explanation as to why the 'vacuum' statement > doesn't > > change 'rowid's when an explicit 'INTEGER PRIMARY KEY' is declared. > > If you declare your own column as being INTEGER PRIMARY KEY, the rowid > will not change by anything SQLite does automatically. This is because > SQLite knows that you know what the column is called, so it thinks you > might be relying on its values. > > The programmer can still change rowid by explicitly updating it, or any > column aliased to it, using an UPDATE command. > > Simon. > _______________________________________________ > sqlite-users mailing list > sqlite-users at mailinglists.sqlite.org > http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users >
