Tom Lane wrote:
> Greg Stark <gsstark(at)mit(dot)edu> writes:
> 
>>Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
>>
>>>How is what you're suggesting more portable?
> 
> 
>>Well, the driver would be free to implement $sth->last_insert_id() using
>>whatever proprietary extensions it has available. The non-portableness would
>>at least be hidden in the driver layer.
> 
> 
> Are you asserting that last_insert_id() is a portable function?  I doubt
> it.

I'm not familiar with the Perl interface, but JDBC has a standardized 
interface for this:

http://java.sun.com/j2se/1.5.0/docs/api/java/sql/Statement.html#executeUpdate(java.lang.String,%20int)
http://java.sun.com/j2se/1.5.0/docs/api/java/sql/Statement.html#getGeneratedKeys()

I tend to agree that a protocol-level change is easier to support in a 
driver. If it's done by extending INSERT/UPDATE, the driver will need to 
parse and modify queries which is hairy at the best of times.

-O

In response to

• Re: [NOVICE] Last ID Problem at 2005-02-02 17:05:22 from Tom Lane

pgsql-hackers by date

Next:	From: Tom Lane	Date: 2005-02-02 21:04:27
	Subject: Re: [NOVICE] Last ID Problem
Previous:	From: Martin Pitt	Date: 2005-02-02 20:42:28
	Subject: Re: libpq API incompatibility between 7.4 and 8.0