Re: ask for review of MERGE

Greg Smith <greg@2ndquadrant.com> writes:
> Marko Tiikkaja wrote:
>> What's been bothering me is that so far there has not been an 
>> agreement on whether we need to protect against concurrency issues or 
>> not.  In fact, there has been almost no discussion about the 
>> concurrency issues which AIUI have been the biggest single reason we 
>> don't have MERGE already.

> Sure there were, they just happened a long time ago.  I tried to 
> summarize the previous discussion at 
> http://wiki.postgresql.org/wiki/SQL_MERGE with the following:

> "PostgreSQL doesn't have a good way to lock access to a key value that 
> doesn't exist yet--what other databases call key range 
> locking...Improvements to the index implementation are needed to allow 
> this feature."

This seems to be presuming there's a consensus that that's the way to
implement it, which is news to me.  Why wouldn't we try the INSERT
first, and if we get a unique-key failure, do UPDATE instead?  I am not
at all in agreement that we ought to build key range locking, nor that
it'd be particularly helpful for this problem even if we had it.
        regards, tom lane