Can somebodey point me in the right direction to find information
regarding Createing a comma delimited file from a universe database
using pick. I am a newbie in regards to Pick/universe so any
references would help... includeing books or tutorials. I am also
relatively new to unix/linux.

Thank You in advance

Tab delimited is a better way to go.

On 8 Jul 2003 14:22:43 -0700, fillheimer (AT) hotmail (DOT) com (Phil) wrote:

phil,

there are many different approaches to this. i have written a generic
utility program that will create .csv files from any universe
datasource. the program prompts you for a source filename and then for
each dictionary id (field) you want to include in the .csv output
file. you simply enter in each dictionary id in the exact order you
want to have the fields for each 'row' in your .csv file layed out and
bada-boom you get a .csv file created. the only drawback is that my
utility *does not* handle multivalued fields. i can email you the
source code. if you are interested, email me.

Ital

Mainly for the original poster:

And someone mentioned briefly that tab delimited is better. Why? Because
tab (char-9) is much less likely to exist in a data file than a comma so it
takes less logic to make the delimited file. That is, you don't need to
escape the commas. You still should escape the char-9 just in case some
prior programmer did something "smart". The general answer to your question
is to write a simple Basic program. If you know a little something about
the file that you are extracting then the program is usually fairly simple,
but if you try to write a general extractor - which many have pursued - then
you will find that is probably almost never ending. Why again? Because the
controlling/dependent relationships can get rather complicated, especially
if there are values and subvalues. If the extracted file is to be read and
used in the flat world then a lot of logic may be needed to get the extra
dimension flattened out without losing anything. Some might advise you to
use ODBC, but you will have the same problem to solve - how to explode out
the values and subvalues that are involved in a non-documented relationship.
Non-documented might be a little strong. Most Pick programmers depend on
the Basic program to tell them about the files rather than the dictionaries
when they are working with inherited systems. Why again, again? Because
there is no requirement to make the dictionary match the data and as a
matter of fact it is rather unusual when that really happens.
If, as I suspect from the wording of the original post, you (the OP) are
not a Pickie then you might say where you are located and someone on the
group will probably contact you with an offer of some kind.
BobJ
"Ital" <I-Tal (AT) mindspring (DOT) com> wrote

Phil:
If you are using one of the many multi-value terminal emulators they
probably have an import function which will except CSV, fixed length or many
other formats. We use Viaduct which handles this very well but I am sure
that Accuterm or most of the others will also perform the same function.
Otherwise, if you want a Pick solution, as BobJ pointed out you could write
a small PICK program to do the same thing.
Bob Marik

"Phil" <fillheimer (AT) hotmail (DOT) com> wrote

I suspect that Phil is wanting to suck the data out of a Universe
Application and put it into something he's familiar with. Spend the money,
Phil, to get someone to help you out. You'll save yourself, and your
client, a LOT of money.

"Robert Marik" <rmarik (AT) mcbaincamera (DOT) com> wrote

There are a number of ways to do this.

There is a program here: ftp://ftp.cedarville.edu/download/download602/
that will create files in many different formats. Simply pick the version
that you need.

hth
Colin
Calgary, Alberta Canada

"Phil" <fillheimer (AT) hotmail (DOT) com> wrote

Sorry forgot the link to the U2 (UniData and UniVerse) documentation.

http://www-3.ibm.com/software/data/u2/pubs/library/

Colin

fillheimer (AT) hotmail (DOT) com (Phil) wrote in message news:<8c0ce8e8.0307081322.13dbb777 (AT) posting (DOT) google.com>...
if you send me some information about your database i'll write a
specific program for it gratis

or you can have a generic solution

daniel_fisher (AT) postmaster (DOT) co.uk

A way I use with great effectiveness (sp?) is to create a dictionary item
called comma
<1> A
<2> 0
<3>
<4>
<5>
<6>
<7>
<8> F;","
<9> L
<10> 1

then I set the printer to record to the &HOLD& file (this system file is a
Type 1 file, meaning that it is a directory at o/s level). I can't think of
the exact syntax, but something like this...

(the 5000 stops the thing from asking for Next Page - literally sets page
length to 5000 lines)

Finally I list the file I require, inserting a COMMA between each field. Of
course this doesn't add in the double quotes...

with the setptr above this creates a file in &HOLD&. To get at the file (on
Windows & nix) just navigate to that directory