about clusters

I have had the some frustration when using ordinary indexes in Oracle. When trying to build views on many tables all indexed in a "normal manner" shareing the same key, Oracle some times doesn't use the indexes, but goes into full table scan when looking for one row in the wiew. I have just been using ordinary indexes, not clusters. I would think that my key was well cut for clusters but I'm told that clusters are only good for static data.

The argument was that if a (cluster) index node pointed out to row in table and the row got updated in such a way that an elment which had been NULL was nolonger NULL, (that is the row got bigger) then what hapened was this: the row had to be removed, inserted somewhere else, and pointer inserted at the orginal place to point to the new place.

And so on.

I'm looking for some people with experience with clusters, if you could tell me about your experiance, that would be great.

Are clusters part of the (draft?) ISO standard? Received on Mon Feb 24 1992 - 00:32:22 CET