Re: explain analyze functionality
От | Tom Lane |
---|---|
Тема | Re: explain analyze functionality |
Дата | |
Msg-id | 28917.1032198878@sss.pgh.pa.us обсуждение исходный текст |
Ответ на | explain analyze functionality (Tourtounis Sotiris <tourtoun@csd.uoc.gr>) |
Список | pgsql-general |
Tourtounis Sotiris <tourtoun@csd.uoc.gr> writes: > i have an SQL query of the general type > select att0 from table where 1 <= att0 and att0 < 1000000 > to a table of the general format (att0 int4, att1 int4) where there is an > index on att0 and when i run the explain analyze on that question it > returns me the following result: > NOTICE: QUERY PLAN: > Index Scan using index1 on class y (cost=0.00..3662.16 rows=1266 width=4) > (actual time=0.44..4323.38 rows=1000000 loops=1) > Total runtime: 4563.40 msec > doews anyone know each of those two index scans, which are stated there > are unique to which of the two conditions refer: to condition 1 <= att0, > to condition att0 < 1000000 or to the union of those two questions? It should be using both. As of 7.3 it's possible to tell this from the EXPLAIN output: regression=# explain select * from tenk1 where 1 < unique1 and unique1 < 100; QUERY PLAN --------------------------------------------------------------------------------- Index Scan using tenk1_unique1 on tenk1 (cost=0.00..353.69 rows=100 width=244) Index Cond: ((1 < unique1) AND (unique1 < 100)) (2 rows) but in older releases you have to grovel through EXPLAIN VERBOSE output if you want to be sure. regards, tom lane
В списке pgsql-general по дате отправления: