I believe it does what you did.
Printable View
I believe it does what you did.
Chained Rows are a fact of life.
Migrated Rows are an application originated issue.
Problem is, Oracle shows both Chained and Migrated Rows as a single count under the Chained Rows label.
In our shop we sound an alarm when "Chained Rows" go over 3% of the table population.
Man, chained and migrated rows are different story, we speak about the free space under the HWM
Poster introduced chained rows issue in his last post.Quote:
Originally Posted by Bore
I consider HWM issue already closed
:)
Oh no, PAVB, I was quoting the manual, and it speak about row-chaining besides the space points in the Segment Advisor´s section, but in the post I say "row chaining, of course, don´t matters in the context what we are talking", space management is the topic here...
SO, Bore, ok, according you the Segment Advisor does the same as me, ok - no other ways to know the list of blocks without data inside a extent ? I will try the advisor, but my concern is about the algorithm, much probably it is a "inside", hidden-one ....
Regards,
Chiappa
u can calculate average block space usage using stats available.. u can get hundreds of such queries on net..
rgds
abhay.
May I ask what are you planning to do with your "list of blocks without data"?Quote:
Originally Posted by JChiappa
i think she already said that in post# 1 (not explictly though).. if i apply a bit of my logic, then i think she wants to reclaim the huge chunk of space thats left unused (or its going to be used in a very long run, but q is why now when it can grow gradually in the long run)
I see... mmhhh... now you tell me how a "list of empty blocks" would help? tell me what are you gonna do with that list.Quote:
Originally Posted by abhaysk
By the way, are you sure poster is a "she"?
PAVB, the idea was : the total number of blocks in a extent is know, if we could easily have the number of blocks without data in the extent, I would calculate a porcent of effective utilization, tables with many extents in a condition of low percent use are in this "white space" condition... But ok, if we don´t have the info easily forget, I will continue to make comparision between number of rows in the table versus total blocks, tables with low number or rows and high number of blocks allocated are the ones.... And of course, "what to do" is simple, re-allocation of the full segment, rebuilding it (with MOVE/REBUILD, reallocation package, or alikes), afaik 10g yet do not let we reallocate/rebuild extents only...
Incidentally you are right, I´m not a "she", I´m male - probably abhaysk saw my name ending with an "a" and assumed it (like Paula, Sheila, Antonia, Shirona), but no, wrong guess...
Regards,
Chiappa