Brian: Your use of the "F" style of correlative begs a question. Did you venture into "A" correlatives? I'm sure you did but I wonder why you offer up an "F" style here. Perhaps a speed consideration?
I first learned "F" style in the Jurrasic Pick era and for 98% of the time, use "A" types today. There are a few "F" methods that don't appear in "A" styles. The "F" style disuaded people from continung with dict items as it was of the Reverse Polish Notation (RPN) style and harder to visualize. Not so many people understood 'stack-style' processing. This is similar to the way HP calculators worked. 4 <enter> 5 <plus> instead of 4 <plus> 5 <equal sign>. Hp calcuators didn't have an <equal sign> nor parenthesis. The "A" style was more readable and was more Data/Basic in readability. Thus it could be more easily maintained. BTW, I've never had any problems with any type of correlative on line 008 regardless of there being an "A" or "S" on line 1. In fact, again I only have one real reason to use an "S" instead of an "A" when creating a new dict item. Here's the "A" correlative in case anyone cares: 008 A;(1["1","4"])(Totherfile;X;1;1) Finally, Tony's suggestion causes the output to be sorted by the [1,4] extraction but display the (Totherfile;X;1;1) value. Not that it's wrong in any way, but using 007 & 008 allows you to break it up in internal (sorted, 008) format then display (converted 007) format. Doing it on a single line, either "F" or "A" style or as an I descriptor doesn't offer the chance to sort by the [1,4] before displaying the translation. Otherwise it sorts by the translation. Just a curiosity. Thanks Mark Johnson ----- Original Message ----- From: "Brian Leach" <[EMAIL PROTECTED]> To: <[email protected]> Sent: Tuesday, May 20, 2008 4:01 AM Subject: RE: [U2] multi-valued extraction > Tony > > A. FYI the ^253 is the representation of a value mark in the UniVerse > editor. So the first example will work correctly. > > B. If I were forced to use a hideous Pick style dictionary rather than a > nice I-descriptor (grin) I would probably use the following, which will also > work successfully: > > F;1;C1;C4;[];(Totherfile;X;1;1) > > But then you get what deserve using A/S types! > > Brian > > > > > > -----Original Message----- > > From: [EMAIL PROTECTED] > > [mailto:[EMAIL PROTECTED] On Behalf Of Tony G > > Sent: 20 May 2008 06:19 > > To: [email protected] > > Subject: RE: [U2] multi-valued extraction > > > > Please have a look at my blog entry: > > remove.pleaseNebula-RnD.com/blog/tech/mv/2007/01/dict-items01.htm > > l > > I explained there how Pick dict items work, where conversions > > and correlatives are processed in a very specific order. I'm > > not familiar enough with the nuances of the Prime influences > > on U2, but this may help a number of us to learn. Aside from > > common documentation, can anyone provide insight into nuances > > about exactly how the U2 platforms (and I really hate to lump > > them together for this) might process descriptors like this? > > > > In D3, one way I'd solve this problem is to use the dict > > attributes as follows: > > > > 001 S > > 002 1 > > 007 Tfile;x;;99 > > 008 T1,4 > > > > Richard, note that your first example of T1,4^TOther.File > > will not work because of the way the dict items are > > processed, as described in my blog. I'm also guessing the > > "253" in there was just a typo. The second and third > > examples are fine for output only, but would be different > > depending on how you'd want the data sorted and selected. > > Again, understanding this for U2 types would be very helpful. > > > > Thanks. > > > > Tony Gravagno > > Nebula Research and Development > > TG@ remove.pleaseNebula-RnD.com > > > > > > > From: Richard Lewis > > > In a Pick-style dictionary, a correlative like this should > > work: > > > > > > T1,4^253TOTHER.FILE;X;;99 > > > or: > > > A1(T1,4)(TOTHER.FILE;X;;99) > > > or: > > > F;1;(T1,4);(TOTHER.FILE;X;;99) > > > > > > or as an I-descriptor expression: > > > > > > TRANS(OTHER.FILE,SUBSTRINGS(F1,1,4),99,'X') > > ------- > > u2-users mailing list > > [email protected] > > To unsubscribe please visit http://listserver.u2ug.org/ > ------- > u2-users mailing list > [email protected] > To unsubscribe please visit http://listserver.u2ug.org/ ------- u2-users mailing list [email protected] To unsubscribe please visit http://listserver.u2ug.org/
