Re: REPLACE INTO table a la mySQL

Christopher Kings-Lynne wrote:
>
> > > I know we're not in the business of copying mySQL,
> > > but the REPLACE INTO table (...) values (...) could be
> > > a useful semantic. This is a combination INSERT or
> > > UPDATE statement. For one thing, it is atomic, and
> > > easier to work with at the application level. Also
> > > if the application doesn't care about previous values,
> > > then execution has fewer locking issues and race
> > > conditions.
> >
> > I don't know if it is standard SQL, but it will save hundreds of
> > lines of code
> > in applications everywhere. I LOVE the idea. I just finished
> > writing a database
> > merge/update program which could have been made much easier to
> > write with this
> > syntax.
>
> The reason MySQL probably has it though is because it doesn't support proper
> transactions.
>
> While we're at it, why not support the MySQL alternate INSERT syntax
> (rehetorical):
>
> INSERT INTO table SET field1='value1', field2='value2';

That is not an issue, but a "REPLACE" syntax can take the place of this:

SQL("select * from table where ID = fubar");

if(HAS_VALUES(SQL))
SQL("update table set xx=yy, www=zz where ID = fubar");
else
SQL("insert into table (...) values (...)");

REPLACE into table set xx=yy, ww = zz where ID = fubar;

A MUCH better solution!