Re: [Hardhats-members] Pointers and keys

[Chris Richardson]

Thurman;      How do you know that the number that you are assigning is unique?  What does the unique number represent? An encounter id which each individual may have a number of these?  Why not let Fileman assign the unique identifier (like it already does for the patient file).  Then the name doesn't need to be the .01 field.  You can always define a cross reference for the field you want to use for the name..  Unfortunately, I can't see how you will keep the unique IDs straight when you supply the name second.  It sounds like you are making more work for yourself.      I hope this helps.   Chris ----- Original Message ----- From: Thurman Pedigo To: hardhats-members@lists.sourceforge.net Sent: Friday, July 29, 2005 8:02 PM Subject: [Hardhats-members] Pointers and keys With so much excitement, I hate to post such a mundane question as pointers and keys.  However, I have never been able to get KEYs to perform in the way I expect. I want to have file TESTFILE point to field .01 of another file so that I can use backward pointers. Further I would like to create a record in the file with a unique # to main the uniqueness of multiple instances of the same name in .01 field of the file. It seems I should be able to enter a unique # such as 12345 without having to encapsulate a new entry as the name file in quotes.  For instance if the name is DOE,JOHN  in the referenced field, I would like to enter 12346 (in field #2) to create a unique entry, then enter DOE,JOHN in field .01, separate from the first record than the 12345 and still allow the .01 field in this record to exist as “DOE,JOHN” . Is this workable? Or is this just bad policy?   I know this is sort of confused and will be difficult to sort out. Basically, I want to have field 2 as the UNIQUE ID for each record in TESTFILE, and .01 field point to a sort of parent file.   Thanks,   thurman