I think that the only reason that it's "more complicated" in later versions is 
that there are so many advances in the later versions that protect our data, in 
particular, the various constraints.

It's actually not that difficult to set up a system to ensure that everything 
is re-loaded in such an order that the constraints can be applied correctly. In 
each of my databases I have a table that lists all the tables with columns that 
show what constraints a particular table relies on and is relied on by. Another 
column is used to manually number the rows so that a cursor can run through 
them in the correct order. Once it's set up it only really changes when new 
tables are added or, more rarely, when constraints are changed.

Unless I have missed a change, R:Base can unload tables in an order that 
prevents an accurate reload if you unload everything in one go. My method takes 
care of this so that I can (quickly) rebuild any database and be sure that 
there are no database errors - ie, RB1 is re-created and the data inserted into 
a completely new set of RBn files. I get the data sorted as a by-product since 
I can unload it in the sequence I prefer.

As for nobody else wanting to have sorted tables they, correctly, say there is 
no "need". I happen to be able to use and like some of my tables pre-sorted - 
particularly those that rarely change. However, my apps take care to sort 
_everything_ and, in any case, the order that an app requires the data may not 
be the same that I can use in the data browser. I think that I used this 
example before but I have a table of music tracks in which the track numbers 
are part of the data. A CD's track sequence is never going to change - track 1 
will never suddenly need to follow track 2 or track 5. However, I may want to 
sort the tracks alphabetically by title and R:Base, of course, does this 
admirably whenever needed. I like to have that table pre-sorted so that the 
tracks for each item are in numeric sequence for viewing in the browser. I see 
that as an additional feature rather than an error of use on my part.

Regards,
Alastair.


  ----- Original Message ----- 
  From: Fred C Kopp 
  To: RBASE-L Mailing List 
  Sent: Wednesday, June 11, 2008 4:46 AM
  Subject: [RBASE-L] - Re: Sorted tables


  Alistair, indeed, you have responded to this issue before and I have looked 
up your message of 5/5/05.  Why can something so simple in 2.11 be so 
complicated in 7.6?  All I want to do is sort my tables and leave my keys 
undisturbed.  Am I to understand that everyone else is content to leave their 
tables unsorted?  If so, I reluctantly bow to the majority, while I continue to 
search for a way to defy it.
    ----- Original Message ----- 
    From: Alastair Burr 
    To: RBASE-L Mailing List 
    Sent: Tuesday, June 10, 2008 6:08 PM
    Subject: [RBASE-L] - Re: Sorted tables


    Looking back, Fred, I seem to think that I, for one, have responded to this 
question before - on the  7th of Feb to be precise!

    I have a command file that re-builds my databases with all the tables 
sorted in the sequence I like - it's yours for the cost of asking for it - it's 
a bit too big to post here.

    As I said previously:

    "I also like to re-build my databases at the turn of each year. As briefly 
discussed here earlier this year I do this by unloading the tables in a 
particular sequence (of tables) so that they can be reloaded to a new database 
without conflicts arising due to missing keys from tables that should have been 
loaded first.

    Since I have to unload the data anyway I do so with the ORDER BY clause 
that best suits each table for my needs.

    This, effectively, kills two birds with one stone: I get a clean database 
and its tables are pre-sorted."

    Regards,
    Alastair.

      ----- Original Message ----- 
      From: Fred C Kopp 
      To: RBASE-L Mailing List 
      Sent: Tuesday, June 10, 2008 8:51 PM
      Subject: [RBASE-L] - Sorted tables


      What if I'm so anal retentive that I want to keep my tables sorted in 
certain ways?

      In 2.11 I simply ran a file that RENAMEd each table to Temp and PROJECTed 
a new, sorted table from Temp.  Then I REMOVEd TABLE Temp and moved on to the 
next table.  After all the sorted tables were projected, my file would BUILD 
KEY for each appropriate column in each appropriate table.  A PACK or RELOAD 
removed the wasted space.

      But things have changed...

      PKs and other Restraints prohibit this naive approach.  I don't even want 
to get into the errors I get.  I want a single control file that sorts my 
tables and preserves / restores my Keys and Restraints.  Doesn't seem all that 
unusual.  How do I do it?

      Thanks,
      Fred

      Fred C. Kopp
      Authorized R:Base Developer
      19 Teri Lane
      Washington, PA  15301

      P 724-222-7376
      F 724-222-7376
      C 724-413-5534
      E [EMAIL PROTECTED]


--------------------------------------------------------------------------


      No virus found in this incoming message.
      Checked by AVG. 
      Version: 7.5.524 / Virus Database: 270.2.0/1493 - Release Date: 
09/06/2008 17:25



------------------------------------------------------------------------------


  No virus found in this incoming message.
  Checked by AVG. 
  Version: 7.5.524 / Virus Database: 270.2.0/1493 - Release Date: 09/06/2008 
17:25

Reply via email to