Tom Lane wrote:
Matt Newell <[EMAIL PROTECTED]> writes:
BTW, i'm on the list now, so no need to cc me.
Common practice around here is to cc people anyway --- this has grown
out of a history of occasionally-slow list mail delivery. If you don't
want it, best to fix it in your mail f
Christopher Kings-Lynne <[EMAIL PROTECTED]> writes:
> Maybe the solution is to make inherited tables actually the same table,
> and jank it with an extra per-row attribute to differentiate them or
> something :)
Aside from destroying the inheritance-for-partitioning stuff, this
wouldn't work for
Most of the people who have thought about this have figured that the
right solution involves a single index spanning multiple tables (hence,
adding a table ID to the index entry headers in such indexes). This
fixes the lookup and entry problems, but it's not any help for the
lock-against-schema-m
Matt Newell <[EMAIL PROTECTED]> writes:
> On Friday 11 November 2005 11:07, you wrote:
>> 1. How do you avoid deadlock among multiple processes all doing the
>> above for similar (same page anyway) keys?
> Isn't all that is required is that they iterate through the indexes in the
> same order.
Y
On Friday 11 November 2005 11:07, you wrote:
> It's an idea, but you are now staring directly into the hornet's nest:
>
> 1. How do you avoid deadlock among multiple processes all doing the
>above for similar (same page anyway) keys? It's difficult if not
>impossible to ensure that they'l
Matt Newell <[EMAIL PROTECTED]> writes:
> Would it be possible to make another routine that locates and aquires
> a write lock on the page where the key would be inserted in each
> index(for each table in the inheritance), and holds all these locks
> until the key is inserted into the correct index
On Thursday 10 November 2005 15:58, you wrote:
> >> The multi-table-unique-constraint problem has to
> >> be solved before we can even think much about multi-table FKs :-(
> >
> > Do you have ideas about how this should be solved?
>
> There's some discussions in the pghackers archives --- look for
