On Wed, 6 Jun 2018 14:33:12 +0300 Aydin Ozgur Yagmur <ayagmu...@gmail.com> escribió:
> I was applying the changes. After moving blob to last column suggested by > Hick and eduardo, 500 MB delete took 40-45 seconds. > I was already attempted to change page_size & cache size but i had not seen > any significant improvement, so i didnt apply this now. When you change the page_size you must vacuum the db. If not, the change in page_size is noop. The change in cache size will speed up the access to blobs, in a basic way. There are better ways to do so but bit more complex. For example, a 4GB cache_size will put, lazily, all db file in memory. > could you please explain positive effects of "autovacuum=incremental" > solution? I did not see too much positive comment on it. Autovacumm incremental, adds sqlite metadata in pages, performing better read access to blobs bigger than your page_size. The trick is never call the pragma incremental_autovacuum. That's, you set incremental but never use it. > Thanks for your help, -- Eduardo <eduardo.mor...@mobelservices.com> _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users