Re: Clustered indexes - again
Posted in 1996
On Wed, 13 Mar 1996, tolly brumbelow wrote: } robert@clark.netkonect.co.uk (Robert Clark) wrote: } } >OK, } } >I have now had a few responses to my original Q about clustered } >indexes - thanks to all of you. } } >Dan Madvig (d-madvig@bethel.edu) said: } >> If you leave it chustered, then every time the index changes } >> the engine has to shuffle the data to match. If the index } >> doesn't change much, it couldn't hurt. } } >Craig (craig.chemistry.chem.utah.edu) said: } >> Creating a clustered index on a table does not affect future } >> inserts or updates. That is to say, new inserts/updates ignore } >> the cluster flag. So, no performance hits. (I think the only } >> reason the cluster flag remains set is so you get an error if } >> you try to cluster on another index.) } } >So given these responses I am confused. Does leaving an index } >clustered effect future inserts/updates etc? } } >Robert Clark } } } > } >-------------------------------------------------------------- } >e-mail: robert@clark.netkonect.co.uk } >snail: 12 Balvernie Grove, Wandsworth, LONDON, SW18 5RU, UK. } >-------------------------------------------------------------- } Informix does not maintane clustering after the index build. } Therefore, a normal insert or update is treated the same } as it would for any other index. About the only reason } to use a clustered index on a high update table in informix } is to alter index to cluster to regain space without unloading } and reloading the table. } } -- } Tolly Brumbelow, Informix DBA } email:tolly.brumbelow@bridge.bellsouth.com } Tolly, What if you have a lookup table? Wouldn't you want to cluster index on that to move things along? Yours, Nick ********************************* Nick Nobbe, Library of Congress NLS/BPH mail: nnob@loc.gov *********************************