How do you specify a files columns/fields? I can create files using
the CREATE.FILE command but only the @ID field is created by doing
this. How can I add new fields and give them a name?

derek54h (AT) gmail (DOT) com wrote:
Derek:

It's as simple as just doing an edit of the dictionary and giving it a
name. Here is sample below to add address field to file called
CLIENTS. Hit return twice at bottom and then type FI to file.

This is simple example below. Date and numeric trabnslations would go
in attribute 7, etc for date and amount fields.

Do you have doumentation on DICTS at all?


P.S. In your programming you do not actually need your DICT setup but
it's preferred for proper programming and docuemtation, but in software
simply write to CLIENTS<1>, CLIENTS<2>, CLIENTS<3>, etc. to file the
data you want into those field in this example file..


----: I
0001= A
0002= 1
0003=
0004=
0005= S
0006=
0007=
0008=
0009= L
0010= 35
Bottom at line 10.
----:

JJ wrote:

http://home.iprimus.com.au/raywurlod...ctionaries.htm

Since only the @ID is required by a the file, that is all it can create for
you. You must define every other field (attribute) by editing a dictionary
entry for the file. The dictionary does not have to be exact but entries
should exist if you want to work with the data. JJ gave you a good
framework. The dictionary, if it is correct (and it does not have to be),
may open vistas for you. You will be amazed what a dictionary entry can do
for you (see virtual dictionary entries).

The next step, indexes. This opens the window to amazing speed.

Try to remember this...the learning curve may be steep but when you approach
the top you will see the world through different eyes.

Wolf.

<derek54h (AT) gmail (DOT) com> wrote

Derek,

You might want to also check the UniVerse manual because the UniVerse and
UniData dictionaries have a slightly different look to them than what JJ
told you. I think what JJ told is correct for other favorites of PICK, but
I'm not sure if it will work for UniVerse, because I have never seen a
UniVerse (or UniData) dictionary take on what JJ wrote to you.

<If it does, I sure someone in this group will correct me.>

Having said that here's how the dictionary entry would look like in UniVerse
or UniData...

DICT<1> -- "D", "I", "V" (for UniData), "X" or "PH"
DICT<2> -- position in the dictionary
DICT<3> -- output masks like "MD2" or "D4/"
DICT<4> -- heading/column name on report
DICT<5> -- length of output. E.g. "10L", "5R"
DICT<6> -- "S" for single value or "M" for multivalue

So, if I was creating a dictionary for INVOICES,


----: I
0001= D
0002= 1
0003= D4/
0004= Invoice^253 Date
0005= 10L
0006= S
Bottom at line 6.
FI

Like I stated, look over the manual and I think this will give you some
better guidance. Also, I would suggest look at an existing dictionary to
give you more insight.

Hope this helps! Good Luck, Good Hunting!

-- Grant


<derek54h (AT) gmail (DOT) com> wrote

Grant W. Boice, Jr. wrote:

Grant and Derek:

Yes there are PRIME style and PICK style dictionaries and both can be
utilized in Universe as we have both in use here. I assume Unidata
works same way as well.

Both styles are covered in the link I sent to Derek..

The descriptions you use above as examples appear to be PRIME style and
the one I sent is PICK style..

I do not believe it matters which he uses as programmers here use both
styles and whatever they are comfortable with is what they stick with.

Enjoy!, Jay

Thanks for the info JJ. I completely forgot PRIME. (Silly me!) I have
not see this type of format in UniData. I try it as an experiment for
myself.

"JJ" <readyjj (AT) yahoo (DOT) com> wrote