On Sun, Jul 31, 2016 at 10:05 AM, Tom Lane wrote:
> After looking at that a bit, I'm strongly tempted to just take out
> the last two sentences of the para, reducing it to the advice concerning
> maintenance_work_mem. That seems sufficient to describe the current
> behavior,
I wrote:
> Peter Geoghegan writes:
>> The CREATE INDEX documentation states:
>> "For hash indexes, the value of effective_cache_size is also relevant
>> to index creation time: PostgreSQL will use one of two different hash
>> index creation methods depending on whether the
Peter Geoghegan writes:
> The CREATE INDEX documentation states:
> "For hash indexes, the value of effective_cache_size is also relevant
> to index creation time: PostgreSQL will use one of two different hash
> index creation methods depending on whether the estimated index size
The CREATE INDEX documentation states:
"For hash indexes, the value of effective_cache_size is also relevant
to index creation time: PostgreSQL will use one of two different hash
index creation methods depending on whether the estimated index size
is more or less than effective_cache_size. For