David,

I stated "The first data entry might access the first record, the next the
1,000th
record, then next on the 5,000th record, and so on and so on. So, unless the
entire table is cached, does it really help? "

Your comment: Yes, because every modern database use b-tree traversal of
indexes to locate records?


My response: So what if the modern database uses b-tree traversal of
indexes to locate the records. What does that have to do with physically
reading the record from the disk on a server?

I can see where the speed of determining the location of the particular
record would be assisted by this, but knowing where the record is and getting
the record are two totally different things. Once the disk address / page /
whatever is deteremine the db engine still has to go to disk to get the
record unless the entire table is in cach or in memeory on the local machine.

So once again, how does all this caching and b-tree traversal speed up the
physical reading of a record that is not in memory. the database engine still
has to go to disk or worse yet - over the network to get the desired record.

If I've got it wrong (which I may well have), please explain where I missing
the poing.

Thanks,

Dennis