Re: Problem query
От | CS DBA |
---|---|
Тема | Re: Problem query |
Дата | |
Msg-id | 4DE6BCB1.3010706@consistentstate.com обсуждение исходный текст |
Ответ на | Re: Problem query (Merlin Moncure <mmoncure@gmail.com>) |
Список | pgsql-performance |
On 06/01/2011 03:15 PM, Merlin Moncure wrote: > On Wed, Jun 1, 2011 at 3:14 PM, CS DBA<cs_dba@consistentstate.com> wrote: >> Hi All; >> >> We have a table with approx 200 columns. about a dozen columns are text data >> types and the rest are a mix of integers , bigint's and double precision >> types. >> >> The table has about 25million rows. >> >> >> The app wants to run a query like this: >> >> select count(pri_num) from max_xtrv_st_t >> where pri_num in (select max(pri_num) from max_xtrv_st_t where 1=1 >> group by tds_cx_ind, cxs_ind_2 ) >> >> I've tried to split the query up but made little progress, pri_num and >> tds_cx_ind are bigint's and cxs_ind_2 is an integer >> >> The table has an index on all 3 columns (3 separate indexes) >> >> Anyone have any thoughts on tuning this query? > need postgres version# and the current explain analyze (or explain, if > you can't wait for it) > > merlin Postgresql version 8.4.2 Explain: QUERY PLAN ------------------------------------------------------------------------------------------------ Aggregate (cost=6551481.85..6551481.86 rows=1 width=8) -> Nested Loop (cost=6550474.85..6551481.35 rows=200 width=8) -> HashAggregate (cost=6550474.85..6550476.85 rows=200 width=8) -> GroupAggregate (cost=5918263.18..6334840.58 rows=17250742 width=20) -> Sort (cost=5918263.18..5968498.96 rows=20094312 width=20) Sort Key: tds_cx_ind, cxs_ind_2 -> Seq Scan on max_xtrv_st_t (cost=0.00..3068701.12 rows=20094312 width=20) -> Index Scan using max_xtrv_st_t_pkey on max_xtrv_st_t (cost=0.00..5.01 rows=1 width=8) Index Cond: (max_xtrv_st_t.pri_num = (max(max_xtrv_st_t.pri_num))) (9 rows) -- --------------------------------------------- Kevin Kempter - Constent State A PostgreSQL Professional Services Company www.consistentstate.com ---------------------------------------------
В списке pgsql-performance по дате отправления: