Re: Dynamic Partitioning using Segment Visibility Maps

Поиск
Список
Период
Сортировка
От Ron Mayer
Тема Re: Dynamic Partitioning using Segment Visibility Maps
Дата
Msg-id 4785183B.2050709@cheapcomplexdevices.com
обсуждение исходный текст
Ответ на Re: Dynamic Partitioning using Segment Visibility Maps  (Chris Browne <cbbrowne@acm.org>)
Ответы Re: Dynamic Partitioning using Segment Visibility Maps  (Ron Mayer <rm_pg@cheapcomplexdevices.com>)
Список pgsql-hackers
Chris Browne wrote:
> _On The Other Hand_, there will be attributes that are *NOT* set in a
> more-or-less chronological order, and Segment Exclusion will be pretty
> useless for these attributes.  

Really?    I was hoping that it'd be useful for any data
with long runs of the same value repeated - regardless of ordering.

My biggest tables are clustered by zip/postal-code -- which means that
while the City, State, Country attributes aren't monotonically increasing
or decreasing; they are grouped tightly together.   I'd expect all queries
for San Francisco to be able to come from at most 2 segments; and all queries
for Texas to be able to come from only a fraction of the whole.


If the segment sizes are configurable - I imagine this would even
be useful for other data - like a people table organized
by last_name,first_name.   "John"'s may be scattered through out
the table -- but at least the John Smith's would all be on one
segment, while the Aaron-through-Jim Smith segments might get excluded.

Or am I missing something?


В списке pgsql-hackers по дате отправления:

Предыдущее
От: "Tim Yardley"
Дата:
Сообщение: tzdata issue on cross-compiled postgresql
Следующее
От: Tom Lane
Дата:
Сообщение: Re: OUTER JOIN performance regression remains in 8.3beta4