On 5/6/16, Richard Hipp <drh at sqlite.org> wrote: > On 5/6/16, Simon Slavin <slavins at bigfraud.org> wrote: >> >> Believe it or not, the fastest way to synchronise the databases is not to >> synchronise the databases. Instead you keep a log of the instructions >> used >> to modify the database. > > Or, this might be an even better solution. Note that the > sqlite3_trace() function (see > https://www.sqlite.org/c3ref/profile.html) can help here. > > A third solution would be to use the session extension, which will be > merged to trunk on the next release. See > https://www.sqlite.org/c3ref/profile.html
I bungled the copy/paste of that URL. Should be https://www.sqlite.org/draft/sessionintro.html > for details. One advantage > of using the session extension is that it will work even if the two > database files are not identical at the onset. A disadvantage is that > the session extension only works with rowid tables, not WITHOUT ROWID > tables. > > -- > D. Richard Hipp > drh at sqlite.org > -- D. Richard Hipp drh at sqlite.org