Rosemary Alles <al...@ipac.caltech.edu> wrote: > 1) If I were to bundle several thousand SELECT statements in a single > transaction - why would it not run faster?
Why do you believe it should? > 2) This is precisely the problem though - each of those statements > will yield rows of results to be parsed with > sqlite3_column - in the context of the user's (my) program. If many > SELECT statements are issued within the context > of a single transaction (repeatedly), how does one deal with the > results without a callback (if using sql_step)? One calls sqlite3_column_*, once for each column, to retrieve values from the current row after each sqlite3_step call. Which is precisely what sqlite3_exec does internally, right after calling sqlite3_step and right before calling your callback. > Yes, > sql_exec is touted to be a wrapper around sql_prepare, bind, step. > However, is does (also - additionally) offer the > option of a user supplied calleback routine which sql_prepare etc. do > not. And that's good because? > Essentially, my question is about context. if many many SELECTS are > bundled in a single transaction using prepare, > bind and step. In what context does one parse the results? Do we not > have synchronizing issue here? Synchronizing between what and what? I must admit I have absolutely no idea what you are talking about. Igor Tandetnik _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
