Re: Clustered Indices

On Mon, 15 May 1995, Cathy Kipp wrote: > The manual may be misleading (I haven't looked at this section recently), > but half of the above is incorrect. When you cluster an index the rows in > your table are physically ordered on disk by that index. If you should add > new rows, change rows, or delete rows, the clustering will not be maintained > (because of the performance implications). So basically you have a good > starting point. Part of the maintenance of creating a clustered table is to > recluster it periodically so that all rows will once again be physically > ordered on disk. > > Regards, > > - Cathy So what happens when a clustered index is created on a dynamic table at the moment of its creation and never again reclustered - which is quite likely if the application developer is not the same person as the DBA? Is there a performance penalty for using a clustered index in this circumstance? Clustered index taking up more disk space? Query optimiser making the wrong guess because rows are not in the expected order? Ideas appreciated. Pete Harris