On Sun, 5 Mar 2017 08:28:44 -0800 Yuri <y...@rawbw.com> wrote: > > I think you?re better off reading the existing BLOB value into > > memory using SELECT, editing it using standard memory-manipulation > > routines, then writing it back with an UPDATE when you?re ready. > > This is very slow.
Can you quantify that? How big is the blob? How big is the edit? How many edits? How fast are N edits to the blob? How fast to read & write the whole blob? How much RAM available to the process? My design question is, why a blob in the first place? If you are making frequent incremental changes to a blob, that suggests the blob has some structure. If you decomposed the blob into rows, perhaps the edit-the-blob problem would go away, and speed/complexity issues with it. --jkl _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
