Thread: pg_statistic, lack of documentation
Hi, http://www.postgresql.org/docs/9.1/interactive/catalog-pg-statistic.html It specifies that entries are created by ANALYZE, but does not mention that if a table is empty the entry for it is not created. Probably it is worth to add to the docs. The test case is below. grayhemp@[local]:5432 test=#create table t1 (i integer);CREATE TABLEgrayhemp@[local]:5432 test=#select stanullfrac, stawidth from pg_statistic where starelid = 't1'::regclass; stanullfrac | stawidth -------------+----------(0 rows) grayhemp@[local]:5432 test=#analyze t1;ANALYZEgrayhemp@[local]:5432 test=#select stanullfrac, stawidth from pg_statistic where starelid = 't1'::regclass; stanullfrac | stawidth -------------+----------(0 rows) grayhemp@[local]:5432 test=#insert into t1 values (1);INSERT 0 1grayhemp@[local]:5432 test=#select stanullfrac, stawidth from pg_statistic where starelid = 't1'::regclass; stanullfrac | stawidth -------------+----------(0 rows) grayhemp@[local]:5432 test=#analyze t1;ANALYZEgrayhemp@[local]:5432 test=#select stanullfrac, stawidth from pg_statistic where starelid = 't1'::regclass; stanullfrac | stawidth -------------+---------- 0 | 4(1 row) -- Sergey Konoplev Blog: http://gray-hemp.blogspot.com LinkedIn: http://ru.linkedin.com/in/grayhemp JID/GTalk: gray.ru@gmail.com Skype: gray-hemp
On Sat, Jan 14, 2012 at 7:34 AM, Sergey Konoplev <gray.ru@gmail.com> wrote: > Hi, > > http://www.postgresql.org/docs/9.1/interactive/catalog-pg-statistic.html > > It specifies that entries are created by ANALYZE, but does not mention > that if a table is empty the entry for it is not created. The actual behavior, in somewhat more detail, is that the existing statistics entries, if any, are retained. I've added a note to that effect to the documentation for ANALYZE, which seems like a more appropriate place than the pg_statistic documentation. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company
On Fri, Jan 27, 2012 at 9:14 PM, Robert Haas <robertmhaas@gmail.com> wrote: > On Sat, Jan 14, 2012 at 7:34 AM, Sergey Konoplev <gray.ru@gmail.com> wrote: > I've added a note to that effect to the documentation for ANALYZE, > which seems like a more appropriate place than the pg_statistic > documentation. Thank you. > > -- > Robert Haas > EnterpriseDB: http://www.enterprisedb.com > The Enterprise PostgreSQL Company -- Sergey Konoplev Blog: http://gray-hemp.blogspot.com LinkedIn: http://ru.linkedin.com/in/grayhemp JID/GTalk: gray.ru@gmail.com Skype: gray-hemp