RE: How accurate is ALL_INDEXES.NUM_ROWS for the PK index?
From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Fri, 16 Jan 2015 11:24:01 +0000
Message-ID: <CE70217733273F49A8A162EE074F64D92826896A_at_EXMBX01.thus.corp>
Date: Fri, 16 Jan 2015 11:24:01 +0000
Message-ID: <CE70217733273F49A8A162EE074F64D92826896A_at_EXMBX01.thus.corp>
I think the absence of the asterisk is a documentation error - note that the sample_size and last_analyzed are also not marked as being dependent on analyze or gather_stats ! Regards Jonathan Lewis http://jonathanlewis.wordpress.com _at_jloracle ________________________________________ From: oracle-l-bounce_at_freelists.org [oracle-l-bounce@freelists.org] on behalf of Thomas Kellerer [thomas.kellerer@mgm-tp.com] Sent: 16 January 2015 08:41 To: oracle-l_at_freelists.org Subject: How accurate is ALL_INDEXES.NUM_ROWS for the PK index? Hello list, as a means to get a quick overview on how many rows each table has, I wonder how accurate the NUM_ROWS column in ALL_INDEXES actually is for the primary key index (or a UNIQUE index on a NOT NULL column). Unlike other columns in that view (and e.g. the NUM_ROWS column in ALL_TABLES) it is not marked as being dependent on gathering statistics. So: when does Oracle update that information and how accurate is it? Does it reflect the number of rows in "near real-time" (within the rules of transactional visibility of course), or is it just as "unreliable" as ALL_TABLES.NUM_ROWS? Regards Thomas -- http://www.freelists.org/webpage/oracle-l --http://www.freelists.org/webpage/oracle-l Received on Fri Jan 16 2015 - 12:24:01 CET