Re: Flat Query
Date: Fri, 14 Oct 2005 13:33:45 -0600
Message-ID: <m3vezzopuu.fsf_at_lhwlinux.garlic.com>
"David Cressey" <david.cressey_at_earthlink.net> writes:
> Actually if you go back 25 years or more, the term referred to files
> that not only were processed sequentially, but also that did not
> contain records within records, or records grouped into record
> groups.
definitely true of the strong tape heritage enforcing sequential access ... however as files from tape started showing up on disks (or dasd ... direct access storage device) ... which could be randomly accessed .... you did start to see sorted files that were being queried using techniques like binary search.
one of the things transition/migration to more strucutre was that update & rebuild of complete file didn't scale. past a certain point the cost of a complete sort & file rebuild was more than the overhead of infrastructure (indexes and other processing) that allowed for incremental updates w/o having to rebuild the complete file every time.
-- Anne & Lynn Wheeler | http://www.garlic.com/~lynn/Received on Fri Oct 14 2005 - 21:33:45 CEST