On Fri, July 20, 2007 2:07 pm, Uber Wannabe wrote: > However, wouldn't obtaining the id in that manner still possibly lead > to > duplication? The DB would take care of the locking on the inserts, > but the > inserts would be populated from a non-locking select, right? Which No. In the case of Oracle (and others) the ID is GENERATED by that first query: $query = "select SEQUENCE_NAME.nextval"; $result = oracle_query($query); $id = oracle_result($result, 0, 0); //this $id is the ID we will use. It is unique. You then have the guaranteed unique ID to pass in for your INSERT: $query = "INSERT INTO whatever (id) VALUES ($id)"; There is no auto_increment on the id field, or it is a pointless auto_increment, as the $id is passed in from the get-go. > Just my thought... I'd stick with the DB's built-in functionality for > that. > As usual, all the information on that is freely available online. The use of a "sequence" is a documented feature in many databases, notably absent in MySQL. Last I checked... MySQL may have added "sequence" while I wasn't watching... The databases with sequences also generally provide a function to return the last-generated new ID on a per connection basis, as well, which is much more similar to MySQL way. -- Some people have a "gift" link here. Know what I want? I want you to buy a CD from some indie artist. http://cdbaby.com/browse/from/lynch Yeah, I get a buck. So? -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
