Kai Wu <[EMAIL PROTECTED]> wrote:
2. The table schema is simple
TABLE MID(id VARCHAR(8),scts VARCHAR(12),daddr VARCHAR(20))
the scenario is it needs first select the "id" by
select id from MID where scts= .. AND daddr=...
after the id field is retrieved, this entry can be deleted by
delete from MID where id=..
or delete from MID where scts=.. AND daddr=..
Apparently "delete" does some repeated work, ie, to locate the
matching entry from the table
(it might be helpful to set id as PRIMARY KEY, but since
insert/delete dominates, its overhead might be higher)
Try
select rowid from MID where scts= .. AND daddr=...
delete from MID where rowid=...
ROWID is a column representing internal B-Tree key. Every table has it,
and every table is "indexed" by it, by nature of the internal data
representation.
Igor Tandetnik
