Simon Riggs <simon@2ndQuadrant.com> writes:
> On Mon, 2008-09-01 at 22:23 +0300, Heikki Linnakangas wrote:
>> Did plan invalidation make it safe to rely on the presence of a unique
>> index for planning decisions?
> My understanding was "Yes" and this case was the specific reason I
> originally wanted to pursue plan invalidation back in 2006.
Yeah, it should work. The theory is that any schema change that could
affect planning should result in broadcasting a relcache inval message
for the table (not just the index, note). I'm pretty confident that
that works for index addition and removal (cf index_update_stats and
index_drop). There might be some situations where we need to force a
relcache inval but don't currently do so --- constraint addition/removal
for instance I'm not too sure about. But that would represent an easily
fixable bug.
regards, tom lane