Very Computer

I need to write a Progress 4GL '.p' script that exports Progress
table records to a CSV (comma-separated value) file, and I have
some questions which the Progress Language Reference Guide doesn't
seem to address.

To refer back to in the questions below, here is the code I will
be using (or something similar), for a table called say 'cust':

  OUTPUT TO cust.d:

  FOR EACH cust:
    EXPORT DELIMITER ',' recid cust.

  OUTPUT CLOSE.

The questions I have are:

  1. Will this 'recid cust' syntax dump for each record its
     'recid' value followed by those of the actual fields
     defined in the schema (and in the same order)? That's
     the intention.

     (I know recids are obsolete; but some tables include fields
     of type recid, which I presume reference records in other
     tables. So the recid values must all be saved, in order
     to maintain all these links. Horrible, but there you go!)

  2. If the table contains array fields, are these written
     as separate comma-separated fields? e.g. if field is
     a 3*array, is it just dumped as if it had been defined
     as three consecutive 'scalar' fields of the same type?

     If not, what is the format? Presumably if not it must
     be a delimited string, and in that case I just need to
     know the delimiter character.

  3. If the table contains date fields, what is the easiest
     way to ensure their values are written in a specified
     format such as dd/mm/yyyy or yyyy-mm-dd?

     (The program must actually dump more than one table.
     But as the same date format would be wanted for all
     tables, a program-wide default would be acceptable.)

     Same kind of thing applies to numeric fields - I don't
     want it 'pretty formatting' them with comma-separated
     thousands!

Many thanks in advance for any replies or pointers.

Cheers

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

---------------------------------------------------------------------------
"Eternity is a long time, especially towards the end."
  Woody Allen
---------------------------------------------------------------------------

See notes below. I am assuming V8 or better here. It is always a good
idea to include your Progress version and OS in any post.

        for each cust:
            put unformatted integer(recid(cust)) ",".
            export delimiter "," cust.
        end.

The fields will be in the order that they are defined in the data
dictionary.

The larger problem of course is that, if you try and restore these
records to a progress database, they will all get new recid's, and your
links will be broken anyway.

Quote:>   2. If the table contains array fields, are these written
>      as separate comma-separated fields? e.g. if field is
>      a 3*array, is it just dumped as if it had been defined
>      as three consecutive 'scalar' fields of the same type?

>      If not, what is the format? Presumably if not it must
>      be a delimited string, and in that case I just need to
>      know the delimiter character.

Quote:>   3. If the table contains date fields, what is the easiest
>      way to ensure their values are written in a specified
>      format such as dd/mm/yyyy or yyyy-mm-dd?

Quote:>      (The program must actually dump more than one table.
>      But as the same date format would be wanted for all
>      tables, a program-wide default would be acceptable.)

>      Same kind of thing applies to numeric fields - I don't
>      want it 'pretty formatting' them with comma-separated
>      thousands!

Hope this helps.

John,

As William stated:

Quote:> The actual format will be whatever is defined in the data dictionary

--
Crittar

Chris Lawery
------------------------------------------------------------------------
Crittar's Profile: http://www.progresstalk.com/member.php?action=getinfo&userid=3271
View this thread: http://www.progresstalk.com/showthread.php?threadid=53177