Re: [ADMIN] weired behavior... after pg_resetxlog-> dump->initdb-->reload.

[Rajesh Kumar Mallah]

Dear Tom, Sorry i could not replicate that issue . But here is something which was the original problem and is replicable. Regds Mallah. rt2=# CREATE UNIQUE INDEX users_1 on users_sample(name); ERROR:  could not create unique index DETAIL:  Table contains duplicated values. rt2=# rt2=# SELECT name from users_sample group by name having count(*) > 1 limit 20; +------+ | name | +------+ +------+ (0 rows) Time: 766.725 ms rt2=# SELECT version(); rt2=# SELECT name from users_sample group by name having count(*) > 1 ; +------+ | name | +------+ +------+ (0 rows) Time: 905.586 ms rt2=# SELECT version(); +----------------------------------------------------------------------------------------------------------+ |                                                 version                                                  | +----------------------------------------------------------------------------------------------------------+ | PostgreSQL 7.4.3 on i686-pc-linux-gnu, compiled by GCC gcc (GCC) 3.2.3 20030502 (Red Hat Linux 3.2.3-34) | +----------------------------------------------------------------------------------------------------------+ (1 row) Time: 338.191 ms Tom Lane wrote: Rajesh Kumar Mallah <[EMAIL PROTECTED]> writes: rt2=# SELECT name from users group by name having count(*) > 1 limit 20; (4 rows) rt2=# SELECT name from users group by name having count(*) > 1 ; (0 rows) Weird. What PG version is this? What does EXPLAIN show for each of these queries (maybe EXPLAIN ANALYZE too)? Can you provide a self-contained test case? regards, tom lane