On Mon, Sep 3, 2012 at 10:07 AM, Ted Roche <[email protected]> wrote: > On Mon, Sep 3, 2012 at 10:01 AM, John Weller <[email protected]> wrote: > >> I worked on a system many years ago using FPW 2.6 which would have benefited >> from this approach! They wanted to store information about the company in a >> table for reference so had a table with a single row and many columns. >> Unfortunately they needed to store more than the 256 field limit so had to >> have 2 tables Company1 and Company2 so we had to remember where each item >> was stored - not good! > > And I worked on a system that went to the opposite extreme: Single > fields were stored in their own rows, in one of four tables: strings, > numbers, dates, booleans. The (multiple) keys were: > > System c(2) > Subsystem c(4) > Module c(4) > Screen c(3) > serial c(2) > > And all of the codes for the keys were stored, yup, in the tables. So, > it had a built-in self-referential data dictionary. While it sounds > unwieldy, the originator (not me!) had designed a series of functions > so that lookups were very efficient. There was a LOT of disk I/O for a > complex screen, but screen drawing on PS/2 50Z's were never that fast > ;) > > And this was FP 2.x, so there were no transactions. A mid-update power > failure or disk error would wreak havoc on data integrity. > > So, a customer address data record would consist of a series of > lookups for Customer Name, Address1, Address2, etc. Adding a field > just meant adding a new record to the data dictionary; no file locks, > no header locks. You had to gracefully deal with NULLish values in > those records until you backfilled, if needed, of course, but the > system had remarkable flexibility. Years later I worked on one of the > OOP databases and was surprised to see a very similar design. > > I think it was extreme as far as RDBMSes was concerned, but it worked. ----------------------
I have one of these at work today for capture of product tests durring manufacture. Some of our bottles have over 250 measurements or potential ones. We have to get averages of samples for a Certificate of Analysis. They use to use views and I have replaced it with SPROCS. Wicked fast pulling from only 4 tables. Granted they have 25 mill rows. -- Stephen Russell Sr. Analyst Ring Container Technology Oakland TN 901.246-0159 cell _______________________________________________ Post Messages to: [email protected] Subscription Maintenance: http://leafe.com/mailman/listinfo/profox OT-free version of this list: http://leafe.com/mailman/listinfo/profoxtech Searchable Archive: http://leafe.com/archives/search/profox This message: http://leafe.com/archives/byMID/profox/CAJidMYLZf+C=zh-k8o-h0-stqqgouvsa1smxkyu7ip1juro...@mail.gmail.com ** All postings, unless explicitly stated otherwise, are the opinions of the author, and do not constitute legal or medical advice. This statement is added to the messages for those lawyers who are too stupid to see the obvious.
