Hi, Relying on sequence will not work (and is a wrong hack) since the use case includes deleting rows explicitly.
I think it's time for a serious simple benchmark with sqlite and say PostgreSQL. PostgeSQL also had performance problems time ago but this has been resolved. Can you describe the hp_table1 schema (CREATE TABLE statement...) and some data (INSERTs)? Yours, S. 2015-01-24 10:33 GMT+01:00 Clemens Ladisch <clem...@ladisch.de>: > Jim Wilcoxson wrote: >> If you have a table where rows are inserted but never deleted, and you >> have a rowid column, you can use this: >> >> select seq from sqlite_sequence where name = 'tablename' > > This works only for an AUTOINCREMENT column. > >> This will return instantly, without scanning any rows or indexes, and >> is much faster than max(rowid) for huge tables. > > Max(rowid) has a special optimization and looks only at the last entry > in the index. It is what SQLite uses internally for tables without > AUTOINCREMENT, and is actually faster than looking up the sequence value > in a separate table. > > > Regards, > Clemens > _______________________________________________ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
